{ "$schema": "https://json-schema.org/draft/2020-12/schema", "_comment": "Ground truth for LetzAI model routing, pricing, and provider endpoints.", "_pricingSchema": { "_description": "Pricing calculation rules", "image": "Use mode key (1k, 2k, hd, 4k, uhd) or _flat for fixed price", "video": "Use _flat OR (_perSecond × duration). Apply multipliers: _audioMultiplier (if audio enabled), _hqMultiplier (if HQ enabled). Multiplier defaults to 1 if not specified. Special: _withAudio overrides flat+audio cost.", "upscaling": "Image: size key (e.g. '1k','2k','4k') or scale-factor key (e.g. '2x','4x','8x') or _flat. Video (mediaType=video): (_perSecond1080p|_perSecond4k) × duration seconds; size 8→1080p, 12→4K." }, "models": { "claude-sonnet-5": { "active": true, "displayName": "Claude Sonnet 5", "description": "Anthropic's latest Sonnet — drop-in upgrade with adaptive thinking, stronger agentic coding, and 1M context. Introductory pricing through Aug 2026.", "dateAdded": "2026-06-30", "type": "text", "provider": "anthropic", "allowBYOK": false, "isDefault": true, "aliases": [ "sonnet-5", "sonnet5", "claude-sonnet-5" ], "capabilities": { "text": { "providerModel": "claude-sonnet-5", "thinkingMode": "adaptive", "effort": "medium", "maxOutputTokens": 128000, "contextEditing": true, "contextWindowTokens": 1000000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": true } } }, "frontend": { "value": "claude-sonnet-5", "label": "Claude Sonnet 5", "description": "Latest Sonnet — adaptive thinking & agents", "providerIcon": "/providers/anthropic.svg", "category": "Anthropic", "showInChatAgent": true, "order": 0 } }, "claude-sonnet-4-6": { "active": false, "displayName": "Claude Sonnet 4.6", "description": "Anthropic's balanced model for everyday chat, planning, and code generation. Fast, reliable tool use with extended thinking.", "dateAdded": "2026-02-01", "type": "text", "provider": "anthropic", "allowBYOK": false, "isDefault": false, "aliases": [ "sonnet", "sonnet-4.6", "claude-sonnet" ], "capabilities": { "text": { "providerModel": "claude-sonnet-4-6", "thinkingMode": "extended", "thinkingBudgetTokens": { "chat": 3000, "canvas": 5000 }, "maxOutputTokens": 64000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": true } } }, "frontend": { "value": "claude-sonnet-4-6", "label": "Claude Sonnet 4.6", "description": "Balanced — everyday chat, planning & code", "providerIcon": "/providers/anthropic.svg", "category": "Anthropic", "showInChatAgent": true, "order": 1 } }, "claude-opus-5": { "active": true, "displayName": "Claude Opus 5", "description": "Anthropic's strongest Opus — step-change over 4.8 for agentic coding, long-horizon work, and professional knowledge tasks. Adaptive thinking; 1M context; near-Fable intelligence at Opus price ($5/$25).", "dateAdded": "2026-07-24", "type": "text", "provider": "anthropic", "allowBYOK": false, "aliases": [ "opus", "opus-5", "opus5", "claude-opus" ], "capabilities": { "text": { "providerModel": "claude-opus-5", "thinkingMode": "adaptive", "maxOutputTokens": 128000, "contextEditing": true, "contextWindowTokens": 1000000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": true } } }, "frontend": { "value": "claude-opus-5", "label": "Claude Opus 5", "description": "Strongest Opus — agents, coding & hard reasoning", "providerIcon": "/providers/anthropic.svg", "category": "Anthropic", "showInChatAgent": true, "order": 2 } }, "claude-opus-4-8": { "active": false, "displayName": "Claude Opus 4.8", "description": "Prior Opus generation — strong reasoning and long-horizon agentic coding with adaptive thinking. Prefer Claude Opus 5 for new work.", "dateAdded": "2026-06-09", "type": "text", "provider": "anthropic", "allowBYOK": false, "aliases": [ "opus-4.8", "opus-4-8", "claude-opus-4.8" ], "capabilities": { "text": { "providerModel": "claude-opus-4-8", "thinkingMode": "adaptive", "maxOutputTokens": 64000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": true } } }, "frontend": { "value": "claude-opus-4-8", "label": "Claude Opus 4.8", "description": "Prior Opus — hard reasoning & big builds", "providerIcon": "/providers/anthropic.svg", "category": "Anthropic", "showInChatAgent": true, "order": 3 } }, "claude-fable-5": { "active": false, "displayName": "Claude Fable 5", "description": "Anthropic's highest-capability widely released model for long-running agents and the hardest work. Adaptive thinking; priced above Opus ($10/$50).", "dateAdded": "2026-06-10", "type": "text", "provider": "anthropic", "allowBYOK": false, "aliases": [ "fable", "fable-5", "claude-fable" ], "capabilities": { "text": { "providerModel": "claude-fable-5", "thinkingMode": "adaptive", "maxOutputTokens": 128000, "contextEditing": true, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": true } } }, "frontend": { "value": "claude-fable-5", "label": "Claude Fable 5", "description": "Highest capability — long-running agents", "providerIcon": "/providers/anthropic.svg", "category": "Anthropic", "showInChatAgent": true, "order": 4 } }, "gpt-5.6-sol": { "active": true, "displayName": "GPT-5.6 Sol", "description": "OpenAI's GPT-5.6 flagship for complex professional work — coding agents, long research, computer use, and multi-tool workflows. Native web search via the Responses API.", "dateAdded": "2026-07-09", "type": "text", "provider": "openai", "allowBYOK": false, "aliases": [ "gpt-5.6", "gpt-5.6-sol", "gpt-sol", "sol" ], "capabilities": { "text": { "providerModel": "gpt-5.6-sol", "thinkingMode": "none", "maxOutputTokens": 128000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": false } } }, "frontend": { "value": "gpt-5.6-sol", "label": "GPT-5.6 Sol", "description": "OpenAI flagship — frontier reasoning & code", "providerIcon": "/providers/chatgpt.png", "category": "OpenAI", "showInChatAgent": true, "order": 6 } }, "gpt-5.5": { "active": true, "displayName": "GPT-5.5", "description": "OpenAI's frontier model for complex reasoning, coding, and agentic work. Native web search via the Responses API.", "dateAdded": "2026-04-24", "type": "text", "provider": "openai", "allowBYOK": false, "aliases": [ "gpt", "gpt-5.5", "openai" ], "capabilities": { "text": { "providerModel": "gpt-5.5", "thinkingMode": "none", "maxOutputTokens": 64000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": false } } }, "frontend": { "value": "gpt-5.5", "label": "GPT-5.5", "description": "OpenAI flagship — reasoning & code", "providerIcon": "/providers/chatgpt.png", "category": "OpenAI", "showInChatAgent": true, "order": 5 } }, "grok-4.6": { "active": true, "displayName": "Grok 4.6", "description": "xAI's newest flagship for coding, agentic tool calling, and long-running visual work. 500k context; reasoning effort low/medium/high/xhigh (default high). Knowledge cutoff February 1, 2026.", "dateAdded": "2026-08-13", "type": "text", "provider": "xai", "allowBYOK": false, "aliases": [ "grok-4.6", "grok-4-6", "grok4.6" ], "capabilities": { "text": { "providerModel": "grok-4.6", "thinkingMode": "none", "effort": "medium", "maxOutputTokens": 128000, "contextWindowTokens": 500000, "supportedFeatures": { "toolUse": true, "vision": true, "webSearch": true, "webFetch": false } } }, "frontend": { "value": "grok-4.6", "label": "Grok 4.6", "description": "xAI flagship — coding, agents & long context", "providerIcon": "/providers/xai.png", "category": "xAI", "showInChatAgent": true, "order": 7 } }, "gemini-3-pro-image": { "active": true, "displayName": "Nano Banana Pro", "description": "Google's state-of-the-art image generation model with improved text rendering, multi-turn editing, and professional-grade controls over lighting, camera, and composition.", "rating": 9.3, "dateAdded": "2025-11-20", "provider": "vertex", "allowBYOK": true, "aliases": [ "gemini", "google/gemini", "gemini-3-pro-image", "gemini-3-pro-image-preview" ], "capabilities": { "generation": { "providerModel": "gemini-3-pro-image", "pricing": { "1k": 80, "2k": 160, "hd": 160, "4k": 240, "uhd": 240 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Built-in safety filter; no external NSFW needed" }, "editing": { "providerModel": "gemini-3-pro-image", "pricing": { "1k": 80, "2k": 160, "hd": 160, "4k": 240, "uhd": 240 }, "defaultMode": "1k", "nsfwCheck": false } }, "fallback": "nano-banana-pro", "_byokNote": "Vertex backend; nbp-inferencesh uses Vertex for frontend" }, "gemini-2.5-flash-image": { "active": false, "displayName": "Gemini 2.5 Flash Image", "description": "Google's fast and cost-effective image generation model with support for 10 aspect ratios and image-only output capabilities.", "dateAdded": "2025-08-26", "provider": "vertex", "allowBYOK": true, "aliases": [ "gemini-2.5-flash-image" ], "capabilities": { "generation": { "providerModel": "gemini-2.5-flash-image", "pricing": { "_flat": 50 }, "nsfwCheck": false, "notes": "Flat pricing regardless of mode" }, "editing": { "providerModel": "gemini-2.5-flash-image", "pricing": { "_flat": 50 }, "nsfwCheck": false } }, "frontend": { "value": "nano-banana", "label": "Nano Banana", "description": "Great Image Editing Model from Google", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/google.webp", "category": "Third-Party", "baseModel": "gemini-2.5-flash-image", "defaultMode": "default", "resolutions": [ "1k" ], "resolutionModes": { "1k": "default" }, "pricingModes": { "1k": "1k" }, "showInGenerate": false, "showInEdit": true, "order": 3 }, "fallback": "nano-banana-pro", "_byokNote": "Routed through InferenceSH when USEVERTEX=FALSE" }, "seedream": { "active": true, "displayName": "Seedream 4.5", "description": "ByteDance's image model with multi-image consistency, excellent text rendering, and up to 4K resolution. Maintains character identity across multiple images.", "rating": 9.0, "dateAdded": "2025-12-04", "provider": "byteplus", "allowBYOK": false, "aliases": [ "seedream", "seedream-4.5", "seedream-4-5", "seedream-4-5-251128", "bytedance-seedream-4.5" ], "capabilities": { "generation": { "providerModel": "ep-20251208175106-nbz5k", "pricing": { "1k": 80, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "2k", "supportsSequentialGeneration": true, "nsfwCheck": true, "notes": "External NSFW check via InferenceSH" }, "editing": { "providerModel": "ep-20251208175106-nbz5k", "pricing": { "1k": 80, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "2k", "nsfwCheck": true } }, "frontend": { "value": "seedream-4-5", "label": "Seedream 4.5", "description": "Powerful Image Generation Model from Bytedance", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/bytedance.png", "category": "Third-Party", "baseModel": "seedream-4-5-251128", "defaultMode": "2k", "resolutions": [ "hd", "4k" ], "resolutionModes": { "hd": "2k", "4k": "4k" }, "showInGenerate": true, "showInEdit": true, "order": 5 } }, "seedream-5": { "active": true, "displayName": "Seedream 5 Pro", "description": "ByteDance's latest Seedream model with Chain-of-Thought reasoning, enhanced prompt following, reference consistency, and up to 4K resolution.", "rating": 9.2, "dateAdded": "2026-07-08", "provider": "byteplus", "allowBYOK": false, "aliases": [ "seedream-5", "seedream5", "seedream-5-0", "seedream-5-0-pro", "dola-seedream-5-0-pro-260628", "seedream-5-0-260128", "bytedance-seedream-5", "bytedance-seedream-5.0", "ep-20260709142316-qqbd2" ], "capabilities": { "generation": { "providerModel": "ep-20260709142316-qqbd2", "pricing": { "1k": 80, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "2k", "supportsSequentialGeneration": false, "nsfwCheck": true, "notes": "External NSFW check via FAL. Seedream 5.0 Pro via ModelArk images API. Omit sequential_image_generation entirely — the API rejects the param even when set to disabled." }, "editing": { "providerModel": "ep-20260709142316-qqbd2", "pricing": { "1k": 80, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "2k", "supportsSequentialGeneration": false, "nsfwCheck": true, "notes": "Omit sequential_image_generation — Seedream 5 Pro rejects the param entirely." } }, "frontend": { "value": "seedream-5", "label": "Seedream 5 Pro", "description": "ByteDance's newest image model — deep reasoning, multi-reference fusion (2–10 images), up to 4K", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "category": "Third-Party", "baseModel": "seedream-5-0-pro", "defaultMode": "2k", "resolutions": [ "hd", "4k" ], "resolutionModes": { "hd": "2k", "4k": "4k" }, "pricingModes": { "hd": "2k", "4k": "4k" }, "showInGenerate": true, "showInEdit": true, "order": 4 } }, "flux-2": { "active": true, "displayName": "Flux 2", "description": "Black Forest Labs' 32B parameter model with multi-reference image support, 4MP editing, improved text rendering, and enhanced photorealism.", "rating": 8.8, "dateAdded": "2025-11-25", "provider": "fal", "allowBYOK": false, "aliases": [ "flux2", "flux-2", "flux-2-max", "flux-2-pro", "flux-2-dev" ], "capabilities": { "generation": { "providerModel": "fal-ai/flux-2-max", "pricing": { "1k": 60, "hd": 120 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "FAL built-in has_nsfw_concepts" }, "editing": { "providerModel": "fal-ai/flux-2-max/edit", "pricing": { "1k": 60, "hd": 120 }, "defaultMode": "1k", "nsfwCheck": false } }, "frontend": { "value": "flux2-max", "label": "Flux2 [max]", "description": "Highest quality generation", "image": "/blog/en-US/image-models/default.png", "providerIcon": "/providers/bfl.jpg", "category": "Third-Party", "baseModel": "flux2", "defaultMode": "1k", "resolutions": [ "1k", "hd" ], "resolutionModes": { "1k": "1k", "hd": "hd" }, "showInGenerate": true, "showInEdit": true, "order": 5 } }, "gpt-image-2": { "active": true, "displayName": "GPT Image 2", "description": "OpenAI's GPT Image 2 via FAL — exceptional typography, fine-detail rendering, and precise mask-based editing up to 4K.", "rating": 9.6, "dateAdded": "2026-04-21", "provider": "fal", "allowBYOK": false, "isDefault": true, "aliases": [ "gpt-image-2", "gpt-image2", "gptimage2", "openai/gpt-image-2" ], "capabilities": { "generation": { "providerModel": "openai/gpt-image-2", "pricing": { "1k": 160, "2k": 240, "hd": 240, "4k": 480, "uhd": 480 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "FAL openai/gpt-image-2; quality locked to 'high'" }, "editing": { "providerModel": "openai/gpt-image-2/edit", "pricing": { "1k": 160, "2k": 240, "hd": 240, "4k": 480, "uhd": 480 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "FAL openai/gpt-image-2/edit; supports image_urls + optional mask_url" } }, "frontend": { "value": "gpt-image-2", "label": "GPT Image 2", "description": "OpenAI's latest image model with exceptional typography", "image": "/blog/en-US/image-models/default.png", "providerIcon": "/providers/chatgpt.png", "providerName": "OpenAI", "category": "Third-Party", "baseModel": "gpt-image-2", "defaultMode": "1k", "resolutions": [ "1k", "hd", "4k" ], "resolutionModes": { "1k": "1k", "hd": "hd", "4k": "4k" }, "showInGenerate": true, "showInEdit": true, "order": 3 } }, "flux-kontext-lora": { "active": false, "displayName": "Flux Kontext LoRA", "description": "Black Forest Labs' in-context image editing model with LoRA support, enabling iterative multi-turn editing with character and style preservation.", "dateAdded": "2025-05-29", "provider": "fal", "allowBYOK": false, "aliases": [], "capabilities": { "editing": { "providerModel": "fal-ai/flux-kontext-lora", "pricing": { "1k": 80, "hd": 160 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Auto-selected when LoRAs are present; not directly requestable by alias" } } }, "_comment_video_edit": "=== VIDEO EDITING MODELS ===", "veo-3.1-extend": { "active": true, "displayName": "Veo 3.1 Extend Video", "description": "Google's video extension model capable of extending existing videos up to 148 seconds with maintained consistency and style.", "dateAdded": "2025-10-15", "provider": "vertex", "allowBYOK": true, "aliases": [ "veo31extend", "veo3.1extend", "veo-3.1-extend", "veo31-extend" ], "capabilities": { "video_editing": { "providerModel": "veo-3.1-generate-001", "pricing": { "_flat": 1500, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "duration": { "min": 8, "max": 8 }, "nsfwCheck": false, "notes": "Video extension via Vertex AI; flat fee per +8s segment (same tiers as video-veo31 generation)" } }, "frontend": { "value": "veo-3.1", "label": "Google VEO 3.1", "providerIcon": "/providers/google.webp", "providerName": "Google", "description": "Video extension model by Google", "supportedModes": [ "extend" ], "order": 2 }, "jobType": "video_edit" }, "kling-o3": { "active": true, "displayName": "Kling O3 Pro Edit", "description": "Kuaishou's unified multimodal video editor with 7-in-1 capabilities including object removal, background swapping, style changes, and character consistency.", "dateAdded": "2026-02-05", "provider": "fal", "allowBYOK": false, "aliases": [ "klingo3", "kling-o3", "kling_o3", "kling-o3-edit" ], "capabilities": { "video_editing": { "providerModel": "fal-ai/kling-video/o3/pro/video-to-video/edit", "pricing": { "_flat": 50 }, "nsfwCheck": false, "notes": "Video-to-video edit; flat pricing" } }, "frontend": { "value": "kling-o3", "label": "Kling O3", "providerIcon": "/providers/kling.png", "providerName": "Kuaishou", "description": "Video editing model by Kling", "supportedModes": [ "edit" ], "order": 1 }, "jobType": "video_edit" }, "beeble-switchx": { "active": true, "displayName": "Beeble SwitchX", "description": "Beeble's video-to-video model that swaps the background or scene in a clip while preserving the original subject, motion, framing and lighting. Not for character or outfit changes. Driven by an optional reference image plus prompt.", "dateAdded": "2026-05-18", "provider": "beeble", "allowBYOK": false, "aliases": [ "beeble", "switchx", "beeble-switchx", "switch-x" ], "capabilities": { "video_editing": { "providerModel": "switchx", "pricing": { "_flat": 50 }, "nsfwCheck": false, "maxFrames": 240, "maxSourcePixels": 2770000, "notes": "Video-to-video background/scene swap only (no character or outfit changes); flat credits same as Kling O3 Pro Edit. Beeble caps: ≤240 frames, ≤2.77M source pixels." } }, "frontend": { "value": "beeble-switchx", "label": "Beeble SwitchX", "providerIcon": "/providers/beeble.png", "providerName": "Beeble", "description": "Background / scene swap by Beeble", "supportedModes": [ "edit" ], "order": 3 }, "jobType": "video_edit" }, "nbp-inferencesh": { "active": true, "displayName": "Nano Banana Pro", "description": "Google Gemini 3 Pro via InferenceSH with improved retries and per-organization API key support.", "rating": 9.3, "dateAdded": "2025-11-20", "provider": "vertex", "allowBYOK": true, "aliases": [ "nbp-inferencesh", "nbp-inf", "nano-banana-pro-inf" ], "capabilities": { "generation": { "providerModel": "gemini-3-pro-image", "providerApp": "google/gemini-3-pro-image", "pricing": { "1k": 80, "2k": 160, "hd": 160, "4k": 240, "uhd": 240 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Gemini 3 Pro via Vertex AI; supports per-org API key override" }, "editing": { "providerModel": "gemini-3-pro-image", "providerApp": "google/gemini-3-pro-image", "pricing": { "1k": 80, "2k": 160, "hd": 160, "4k": 240, "uhd": 240 }, "defaultMode": "1k", "nsfwCheck": false } }, "frontend": { "value": "nano-banana-pro", "label": "Nano Banana Pro", "description": "The latest Image Generation Model from Google", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/google.webp", "category": "Third-Party", "baseModel": "gemini-3-pro-image", "defaultMode": "default", "resolutions": [ "1k", "hd", "4k" ], "resolutionModes": { "1k": "default", "hd": "2k", "4k": "4k" }, "pricingModes": { "1k": "1k", "hd": "2k", "4k": "4k" }, "showInGenerate": true, "showInEdit": true, "order": 1 }, "fallback": "nano-banana-pro", "_byokNote": "Always uses InferenceSH; org key used when orgId is present" }, "gemini-3-1-flash-image": { "active": true, "displayName": "Nano Banana 2", "description": "Google's Gemini 3.1 Flash image generation model — fast, cost-effective generation and editing with up to 4K resolution, advanced text rendering, and Google Search grounding.", "rating": 8.6, "dateAdded": "2026-02-26", "provider": "vertex", "allowBYOK": true, "aliases": [ "gemini-3-1-flash", "gemini-3.1-flash-image", "gemini-3-1-flash-image", "gemini-3-1-flash-image-preview", "gemini-3.1-flash-image-preview" ], "capabilities": { "generation": { "providerModel": "gemini-3.1-flash-image", "pricing": { "1k": 40, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Gemini 3.1 Flash Image (GA) via Vertex AI; built-in safety filter" }, "editing": { "providerModel": "gemini-3.1-flash-image", "pricing": { "1k": 50, "2k": 100, "hd": 100, "4k": 150, "uhd": 150 }, "defaultMode": "1k", "nsfwCheck": false } }, "fallback": "nano-banana-pro", "_byokNote": "Vertex backend; nbf-inferencesh uses InferenceSH for frontend" }, "nbf-inferencesh": { "active": true, "displayName": "Nano Banana 2", "description": "Google Gemini 3.1 Flash via InferenceSH — fast, cost-effective image generation and editing with up to 4K resolution and Google Search grounding.", "rating": 8.6, "dateAdded": "2026-02-26", "provider": "inferencesh", "allowBYOK": true, "aliases": [ "nbf-inferencesh", "nbf-inf", "nano-banana-flash-inf" ], "capabilities": { "generation": { "providerModel": "gemini-3.1-flash-image", "providerApp": "google/gemini-3-1-flash-image", "pricing": { "1k": 40, "2k": 80, "hd": 80, "4k": 160, "uhd": 160 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Gemini 3.1 Flash Image via InferenceSH; Vertex providerModel is GA" }, "editing": { "providerModel": "gemini-3.1-flash-image", "providerApp": "google/gemini-3-1-flash-image", "pricing": { "1k": 50, "2k": 100, "hd": 100, "4k": 150, "uhd": 150 }, "defaultMode": "1k", "nsfwCheck": false } }, "frontend": { "value": "nano-banana-2", "label": "Nano Banana 2", "description": "Fast Image Generation Model from Google", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/google.webp", "category": "Third-Party", "baseModel": "nbf-inferencesh", "defaultMode": "default", "resolutions": [ "1k", "hd", "4k" ], "resolutionModes": { "1k": "default", "hd": "2k", "4k": "4k" }, "pricingModes": { "1k": "1k", "hd": "2k", "4k": "4k" }, "showInGenerate": true, "showInEdit": true, "order": 2 }, "_byokNote": "Uses InferenceSH; org key used when orgId is present" }, "gemini-3-1-flash-lite-image": { "active": true, "displayName": "Nano Banana 2 Lite", "description": "Google's fastest and cheapest Gemini image model — optimized for high-volume generation where speed and cost matter most. Best for single-prompt text-to-image; not optimized for multiple reference inputs or multi-turn editing.", "rating": 8.0, "dateAdded": "2026-06-30", "provider": "vertex", "allowBYOK": true, "aliases": [ "gemini-3-1-flash-lite", "gemini-3.1-flash-lite-image", "gemini-3-1-flash-lite-image", "nano-banana-2-lite" ], "capabilities": { "generation": { "providerModel": "gemini-3.1-flash-lite-image", "pricing": { "1k": 20 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Gemini 3.1 Flash Lite Image via Vertex AI; built-in safety filter" }, "editing": { "providerModel": "gemini-3.1-flash-lite-image", "pricing": { "1k": 25 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Supported but not optimized for multi-reference or sequential edits" } }, "fallback": "nano-banana-pro", "_byokNote": "Vertex backend; nbl-inferencesh uses InferenceSH for frontend" }, "nbl-inferencesh": { "active": true, "hideFromModelsPage": true, "displayName": "Nano Banana 2 Lite", "description": "Google Gemini 3.1 Flash Lite Image via InferenceSH — fastest, most cost-effective image generation for high-volume workloads.", "rating": 8.0, "dateAdded": "2026-06-30", "provider": "inferencesh", "allowBYOK": true, "aliases": [ "nbl-inferencesh", "nbl-inf", "nano-banana-lite-inf" ], "capabilities": { "generation": { "providerModel": "gemini-3.1-flash-lite-image", "providerApp": "google/gemini-3-1-flash-lite-image", "pricing": { "1k": 20 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Gemini 3.1 Flash Lite Image via InferenceSH; supports per-org API key override" }, "editing": { "providerModel": "gemini-3.1-flash-lite-image", "providerApp": "google/gemini-3-1-flash-lite-image", "pricing": { "1k": 25 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Supported but not optimized for multi-reference or sequential edits" } }, "frontend": { "value": "nano-banana-2-lite", "label": "Nano Banana 2 Lite", "description": "Fastest & cheapest image generation from Google", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/google.webp", "category": "Third-Party", "baseModel": "nano-banana-2-lite", "defaultMode": "default", "resolutions": [ "1k" ], "resolutionModes": { "1k": "default" }, "pricingModes": { "1k": "1k" }, "showInGenerate": true, "showInEdit": true, "order": 3 }, "_byokNote": "Uses InferenceSH; org key used when orgId is present. Gemini 3.1 Flash-Lite supports 1K only (2K/4K unsupported by Google)." }, "wan-2-7-image-pro": { "active": true, "displayName": "WAN 2.7 Image Pro", "description": "Alibaba's professional image generation model supporting text-to-image, image editing, and multi-reference generation with up to 4K high-definition output. Supports thinking mode for improved quality.", "rating": 8.4, "dateAdded": "2026-04-01", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "wan27", "wan-2-7", "wan-2.7", "wan-2-7-image-pro", "wan-2.7-image-pro", "wan27-image-pro" ], "capabilities": { "generation": { "providerModel": "wan-2-7-image-pro", "providerApp": "alibaba/wan-2-7-image-pro@7394ycc2", "pricing": { "1k": 10, "2k": 10, "hd": 10, "4k": 10, "uhd": 10 }, "defaultMode": "2k", "nsfwCheck": false, "notes": "WAN 2.7 Image Pro via InferenceSH; supports up to 9 reference images, thinking mode, sequential image sets; up to 4K" }, "editing": { "providerModel": "wan-2-7-image-pro", "providerApp": "alibaba/wan-2-7-image-pro@7394ycc2", "pricing": { "1k": 10, "2k": 10, "hd": 10 }, "defaultMode": "1k", "nsfwCheck": false, "notes": "Same app as generation; reference_images used for editing; max 2K for image-to-image" } }, "frontend": { "value": "wan-2-7", "label": "WAN 2.7 Pro", "description": "Professional Image Generation Model from Alibaba", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/wan.png", "category": "Third-Party", "baseModel": "wan-2-7-image-pro", "defaultMode": "2k", "resolutions": [ "1k", "hd" ], "resolutionModes": { "1k": "1k", "hd": "2k" }, "pricingModes": { "1k": "1k", "hd": "2k" }, "showInGenerate": true, "showInEdit": true, "order": 5 } }, "grok-imagine": { "active": true, "displayName": "Grok Imagine 2.0", "description": "xAI's newest image model — precise instruction following, sharp typography, and iterative editing with up to 5 reference images. $0.04/image; 1K and 2K.", "rating": 9.2, "dateAdded": "2026-08-07", "provider": "xai", "allowBYOK": false, "aliases": [ "grok-imagine", "grok-imagine-2", "grok-imagine-image-2.0", "grok-imagine-image-2", "grok-imagine-image", "grok-imagine-image-pro", "grok", "aurora", "xai" ], "capabilities": { "generation": { "providerModel": "grok-imagine-image-2.0", "pricing": { "_flat": 30 }, "defaultMode": "1k", "nsfwCheck": true, "notes": "xAI Grok Imagine Image 2.0 via REST API; $0.04/image; aspect_ratio, resolution 1k/2k, quality low/medium (default medium)" }, "editing": { "providerModel": "grok-imagine-image-2.0", "pricing": { "_flat": 30 }, "defaultMode": "1k", "nsfwCheck": true, "notes": "xAI Grok Imagine Image 2.0 edit via /images/edits; up to 5 input images; JSON body" } }, "frontend": { "value": "grok-imagine", "label": "Grok Imagine 2.0", "description": "Precise image generation by xAI", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/xai.png", "category": "Third-Party", "baseModel": "grok-imagine-image-2.0", "defaultMode": "1k", "resolutions": [ "1k", "2k" ], "resolutionModes": { "1k": "1k", "2k": "2k" }, "showInGenerate": true, "showInEdit": true, "order": 6 } }, "grok-imagine-quality": { "active": true, "displayName": "Grok Imagine Image Quality", "description": "xAI's quality-focused image model — photorealistic rendering, strong text and logo accuracy, and multi-style versatility from portraits to anime. $0.05/image; 1K and 2K.", "rating": 9.0, "dateAdded": "2026-08-13", "provider": "xai", "allowBYOK": false, "aliases": [ "grok-imagine-quality", "grok-imagine-image-quality", "grok-imagine-image-quality-latest" ], "capabilities": { "generation": { "providerModel": "grok-imagine-image-quality", "pricing": { "_flat": 40 }, "defaultMode": "1k", "nsfwCheck": true, "notes": "xAI Grok Imagine Image Quality via REST API; $0.05/image; aspect_ratio, resolution 1k/2k" }, "editing": { "providerModel": "grok-imagine-image-quality", "pricing": { "_flat": 40 }, "defaultMode": "1k", "nsfwCheck": true, "notes": "xAI Grok Imagine Image Quality edit via /images/edits; up to 3 input images; JSON body" } }, "frontend": { "value": "grok-imagine-quality", "label": "Grok Imagine Image Quality", "description": "Photorealistic image generation by xAI", "image": "/blog/en-US/image-models/cinematic.png", "providerIcon": "/providers/xai.png", "category": "Third-Party", "baseModel": "grok-imagine-image-quality", "defaultMode": "1k", "resolutions": [ "1k", "2k" ], "resolutionModes": { "1k": "1k", "2k": "2k" }, "showInGenerate": true, "showInEdit": true, "order": 7 } }, "_comment_fallback": "=== FALLBACK MODELS ===", "nano-banana-pro": { "active": true, "displayName": "Nano Banana Pro", "description": "Fallback model via Fal.ai used when primary Gemini endpoints return 429 rate limit errors.", "dateAdded": "2025-11-20", "provider": "fal", "allowBYOK": false, "aliases": [], "capabilities": { "generation": { "providerModel": "fal-ai/nano-banana-pro", "editProviderModel": "fal-ai/nano-banana-pro/edit", "pricing": {}, "nsfwCheck": false, "notes": "Gemini 429 fallback only; no direct pricing (inherits from the model it fell back from)" } }, "isFallbackOnly": true }, "_comment_video": "=== VIDEO GENERATION MODELS (resolved by settings.mode) ===", "video-default": { "active": false, "displayName": "Default Video", "description": "Our first generation in-house default video model for video generation.", "dateAdded": "2025-07-01", "provider": "letzai", "allowBYOK": true, "aliases": [ "default" ], "capabilities": { "video": { "providerModel": "infsh/lightning-wan-2-2-i2v-a14b", "providerVersion": "06pqyf4ahanxy68c11b4fwcdba", "pricing": { "_perSecond": 60, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": false, "supportsReference": false, "resolutions": [ "720p", "1080p" ], "duration": { "min": 2, "max": 6 }, "nsfwCheck": true, "notes": "Default video workflow; per-second pricing" } }, "jobType": "video_generation" }, "video-veo3": { "active": false, "displayName": "Google Veo 3.0", "description": "Google's video generation model via Vertex AI with native audio and cinematic quality at fixed 8-second duration.", "dateAdded": "2025-05-20", "provider": "vertex", "allowBYOK": false, "aliases": [ "veo3" ], "capabilities": { "video": { "providerModel": "veo-3.0-generate-001", "providerModelFast": "veo-3.0-fast-generate-001", "pricing": { "_flat": 1500, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": true, "supportsReference": false, "duration": { "min": 8, "max": 8 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "notes": "Google Veo 3.0 via Vertex AI; fixed 8s duration" } }, "jobType": "video_generation" }, "video-veo31": { "active": true, "displayName": "Google VEO3.1", "description": "Google's updated video model via Vertex AI with richer native audio, improved dialogue sync, and enhanced cinematic style.", "dateAdded": "2025-10-15", "provider": "vertex", "allowBYOK": false, "aliases": [ "veo31" ], "capabilities": { "video": { "providerModel": "veo-3.1-generate-001", "providerModelT2V": "veo-3.1-generate-001", "providerModelFast": "veo-3.1-fast-generate-001", "pricing": { "_flat": 1500, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": true, "supportsReference": true, "duration": { "min": 8, "max": 8 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "original", "16:9", "9:16" ], "features": { "multiFrame": true }, "notes": "Google Veo 3.1 via Vertex AI; fixed 8s duration; image-to-video (reference image); first/last frame; aspect ratios 16:9, 9:16, original" } }, "jobType": "video_generation" }, "video-veo31inf": { "active": false, "displayName": "Google VEO3.1", "description": "Google Veo 3.1 via Vertex AI with native audio, improved dialogue sync, and 8-second cinematic video generation.", "dateAdded": "2025-10-15", "provider": "vertex", "allowBYOK": true, "aliases": [ "veo31inf" ], "capabilities": { "video": { "providerModel": "veo-3.1-generate-001", "providerModelFast": "veo-3.1-fast-generate-001", "pricing": { "_flat": 1500, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": true, "supportsReference": true, "duration": { "min": 8, "max": 8 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "original", "16:9", "9:16" ], "notes": "Veo 3.1 via Vertex AI; aspect ratios 16:9, 9:16, original" } }, "jobType": "video_generation" }, "video-sora2": { "active": false, "displayName": "Sora 2", "description": "OpenAI's Sora 2 with custom characters, 16:9/9:16 exports, clips up to 20 seconds, video continuation, and image-to-video support.", "dateAdded": "2025-09-30", "provider": "openai", "allowBYOK": false, "aliases": [ "sora2" ], "capabilities": { "video": { "providerModel": "sora-2", "providerModelT2V": "sora-2", "providerModelPro": "sora-2-pro", "pricing": { "_flat": 100, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": true, "supportsReference": true, "duration": { "min": 4, "max": 20 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "16:9", "9:16" ], "notes": "OpenAI Sora 2 native API; input_reference for i2v; 4/8/12/16/20s; 1280x720, 720x1280, 1920x1080, 1080x1920" } }, "jobType": "video_generation" }, "video-kling25": { "active": false, "displayName": "Kling V2.5", "description": "Kuaishou's turbo video model optimized for faster generation with 5-10 second video support.", "dateAdded": "2025-08-01", "provider": "fal", "allowBYOK": false, "aliases": [ "kling25" ], "capabilities": { "video": { "providerModel": "fal-ai/kling-video/v2.5-turbo/pro/image-to-video", "pricing": { "_flat": 100, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": false, "supportsReference": false, "duration": { "min": 5, "max": 10 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "notes": "Kling V2.5 via Fal.ai" } }, "jobType": "video_generation" }, "video-kling3": { "active": true, "displayName": "Kling V3", "description": "Kuaishou's flagship video model with native multilingual audio, up to 15 seconds, multi-shot storytelling, and reference image support.", "dateAdded": "2026-02-05", "provider": "fal", "allowBYOK": false, "aliases": [ "kling3", "kling30" ], "capabilities": { "video": { "providerModel": "fal-ai/kling-video/v3/pro/image-to-video", "providerModelT2V": "fal-ai/kling-video/v3/pro/text-to-video", "imageArgKey": "start_image_url", "generateAudioArgKey": "generate_audio", "pricing": { "_flat": 150, "_withAudio": 300 }, "allowAudio": true, "supportsReference": true, "supportsElements": true, "duration": { "min": 3, "max": 15 }, "resolutions": [ "1080p" ], "aspectRatios": [ "original", "16:9", "9:16" ], "nsfwCheck": true, "features": { "multiPrompt": true, "multiFrame": true }, "multiPrompt": { "maxDuration": 15, "minShotDuration": 3 }, "notes": "Kling V3 via Fal.ai; supports reference images and elements; 1080p only; aspect ratios 16:9, 9:16, original; special audio pricing (300 credits with audio)" } }, "jobType": "video_generation" }, "video-kling26": { "active": true, "displayName": "Kling V2.6", "description": "Kuaishou's video model with simultaneous audio-visual generation, creating videos with voiceovers and sound effects in a single pass.", "dateAdded": "2025-12-03", "provider": "fal", "allowBYOK": false, "aliases": [ "kling26" ], "capabilities": { "video": { "providerModel": "fal-ai/kling-video/v2.6/pro/image-to-video", "providerModelT2V": "fal-ai/kling-video/v2.6/pro/text-to-video", "pricing": { "_flat": 150, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": true, "supportsReference": false, "duration": { "min": 5, "max": 10, "step": 5 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "original", "16:9", "9:16" ], "notes": "Kling V2.6 via Fal.ai; supports audio and text-to-video; aspect ratios 16:9, 9:16, original" } }, "jobType": "video_generation" }, "video-wan25": { "active": true, "displayName": "WAN 2.5", "description": "Alibaba's open-source video model with multilingual support, synchronized audio, and up to 10 seconds at 1080p resolution.", "dateAdded": "2025-09-24", "provider": "fal", "allowBYOK": false, "aliases": [ "wan25" ], "capabilities": { "video": { "providerModel": "fal-ai/wan-25-preview/image-to-video", "pricing": { "_flat": 110, "_audioMultiplier": 2, "_hqMultiplier": 2 }, "allowAudio": false, "supportsReference": false, "duration": { "min": 5, "max": 10, "step": 5 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "original", "16:9", "9:16" ], "notes": "WAN 2.5 via Fal.ai; image-to-video only; aspect ratios 16:9, 9:16, original" } }, "jobType": "video_generation" }, "video-minimax-h3": { "active": true, "displayName": "MiniMax H3", "description": "MiniMax's frontier video model at 2K. Smart-routes text-to-video, image-to-video (optional last frame), and multimodal reference-to-video (reference images, one video, one audio) while keeping subjects consistent.", "dateAdded": "2026-07-31", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "minimax-h3", "minimaxh3", "minimax", "h3", "hailuo-h3" ], "capabilities": { "video": { "providerModel": "minimax/h3", "providerModelT2V": "minimax/h3", "defaultResolution": "2K", "pricing": { "_perSecond": 210, "_perSecond2K": 210 }, "allowAudio": false, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 11, "images": 9, "videos": 1, "audio": 1 }, "duration": { "min": 5, "max": 15 }, "nsfwCheck": true, "resolutions": [ "2K" ], "aspectRatios": [ "adaptive", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "notes": "MiniMax H3 via InferenceSH (minimax/h3). Single endpoint handles T2V/I2V/R2V via input fields (image/last_image for first+last frame; reference_images + singular reference_video/reference_audio for multimodal — cannot combine refs with first/last frame). Duration 5–10s for T2V/I2V, 5–15s for R2V. 2K only; $0.26/s → 210 credits/s. T2V requires an explicit ratio (not adaptive)." } }, "frontend": { "value": "minimax-h3", "label": "MiniMax H3", "providerIcon": "/providers/minimax.png", "providerName": "MiniMax", "description": "Frontier 2K video with multimodal references", "supportedModes": [ "generate" ], "order": 4 }, "jobType": "video_generation" }, "video-flux3": { "active": true, "displayName": "Flux 3 Video", "description": "Black Forest Labs' Flux 3 Video — generate and animate video up to 20s at HD or Full HD with synchronized audio. Supports text-to-video, image-to-video (start frame), and first+last frame keyframes. No Omni / multimodal reference stack.", "dateAdded": "2026-08-04", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "flux3", "flux-3", "flux-3-video", "flux3-video", "bfl-flux-3", "bfl-flux3" ], "capabilities": { "video": { "providerModel": "bfl/flux-3-video@4pmwnz9p", "providerModelT2V": "bfl/flux-3-video@4pmwnz9p", "defaultResolution": "720p", "pricing": { "_perSecond": 100, "_perSecond720p": 100, "_perSecond1080p": 160 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": false, "duration": { "min": 5, "max": 20 }, "nsfwCheck": true, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "auto", "21:9", "2:1", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "notes": "Flux 3 Video via InferenceSH (bfl/flux-3-video). T2V (prompt), I2V (image as start frame), and first+last frame keyframes only — no Omni / R2V multimodal reference stack. Resolution mapped 720p→hd, 1080p→fhd. Audio included at no extra charge (always on). Pricing: 100 credits/s @720p, 160 credits/s @1080p." } }, "frontend": { "value": "flux3", "label": "Flux 3 Video", "providerIcon": "/providers/bfl.jpg", "providerName": "Black Forest Labs", "description": "HD/FHD video with native audio and first/last-frame control", "supportedModes": [ "generate" ], "order": 5 }, "jobType": "video_generation" }, "video-seedance2": { "active": true, "displayName": "Seedance 2.0", "description": "ByteDance's most advanced video model with cinematic output, native audio, real-world physics, and director-level camera control. Supports up to 4K (10-bit color). Supports text, image, audio, and video reference inputs (up to 9 reference images, 3 reference videos, 3 reference audios) with first+last frame control.", "dateAdded": "2026-04-08", "provider": "byteplus", "allowBYOK": false, "aliases": [ "seedance2", "seedance20", "seedance-2" ], "capabilities": { "video": { "providerModel": "dreamina-seedance-2-0-260128", "providerModelInferenceSH": "bytedance/seedance-2-0", "providerModelT2V": "dreamina-seedance-2-0-260128", "pricing": { "_perSecond": 210, "_perSecond480p": 105, "_perSecond720p": 210, "_perSecond1080p": 500, "_perSecond4k": 1000 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 12, "images": 9, "videos": 3, "audio": 3 }, "duration": { "min": 4, "max": 15 }, "resolutions": [ "480p", "720p", "1080p", "4k" ], "aspectRatios": [ "auto", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "notes": "Seedance 2.0 standard via BytePlus ModelArk (dreamina-seedance-2-0-260128) when USE_BYTEPLUS_SEEDANCE=TRUE, or InferenceSH (bytedance/seedance-2-0) when USE_BYTEPLUS_SEEDANCE=FALSE. No cross-provider retry. Single endpoint handles I2V/T2V/R2V; per-second pricing (1000 credits/s @4k, 500/s @1080p, 210/s @720p, 105/s @480p); audio always included." } }, "frontend": { "value": "seedance2", "label": "Seedance 2.0", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "description": "Cinematic video with native audio", "supportedModes": [ "generate" ], "order": 3 }, "jobType": "video_generation" }, "video-seedance2-fast": { "active": false, "displayName": "Seedance 2.0 Fast", "description": "ByteDance's faster Seedance 2.0 Studio variant. Cinematic output with native audio at up to 720p, optimized for speed. Supports text, image, audio, and video reference inputs (up to 9 reference images, 3 reference videos, 3 reference audios) and first+last frame control.", "dateAdded": "2026-05-13", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "seedance2-fast", "seedance20-fast", "seedance-2-fast", "seedance-2-0-fast", "seedance-2-0-studio-fast" ], "capabilities": { "video": { "providerModel": "bytedance/seedance-2-0-studio-fast", "providerModelT2V": "bytedance/seedance-2-0-studio-fast", "pricing": { "_perSecond": 105, "_perSecond480p": 63, "_perSecond720p": 105 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 12, "images": 9, "videos": 3, "audio": 3 }, "duration": { "min": 4, "max": 15 }, "resolutions": [ "480p", "720p" ], "aspectRatios": [ "auto", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "acceptsSafetyIdentifier": true, "notes": "Seedance 2.0 Fast via InferenceSH (bytedance/seedance-2-0-studio-fast); single endpoint handles I2V/T2V/R2V via input fields (image/end_image for first+last frame; reference_images/reference_videos/reference_audios arrays for multimodal); per-second pricing (105 credits/s @720p, 63/s @480p); audio always included; forwards safety_identifier when an org context is present (optional, not enforced)" } }, "frontend": { "value": "seedance2-fast", "label": "Seedance 2.0 Fast", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "description": "Fast cinematic video with native audio (up to 720p)", "supportedModes": [ "generate" ], "order": 4 }, "jobType": "video_generation" }, "video-seedance2-enterprise": { "active": true, "displayName": "Seedance 2.0 Enterprise", "description": "ByteDance's professional Seedance 2.0 Studio variant with private asset library support. Supports up to 4K (10-bit color), with text-to-video, image-to-video, and multimodal reference-to-video (up to 9 reference images, 3 reference videos, 3 reference audios) and synchronized audio.", "dateAdded": "2026-05-13", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "seedance2-enterprise", "seedance20-enterprise", "seedance-2-enterprise", "seedance-2-0-enterprise", "seedance-2-0-studio", "seedance2-studio" ], "enterpriseOnly": true, "capabilities": { "video": { "providerModel": "bytedance/seedance-2-0-studio", "providerModelT2V": "bytedance/seedance-2-0-studio", "pricing": { "_perSecond": 210, "_perSecond480p": 105, "_perSecond720p": 210, "_perSecond1080p": 500, "_perSecond4k": 1000 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 12, "images": 9, "videos": 3, "audio": 3 }, "duration": { "min": 4, "max": 15 }, "resolutions": [ "480p", "720p", "1080p", "4k" ], "aspectRatios": [ "auto", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "requireSafetyIdentifier": true, "acceptsSafetyIdentifier": true, "notes": "Seedance 2.0 Enterprise via InferenceSH (bytedance/seedance-2-0-studio); single endpoint handles I2V/T2V/R2V via input fields; per-second pricing (1000 credits/s @4k, 500/s @1080p, 210/s @720p, 105/s @480p); audio always included; enterprise only; supports up to 9 reference images, 3 reference videos, 3 reference audios with byteplus virtual portrait library. Requires a valid LetzAI organizationId (validated against /internal/organizations/{id} and forwarded as InferenceSH safety_identifier for per-org asset-library namespacing)." } }, "frontend": { "value": "seedance2-enterprise", "label": "Seedance 2.0 Enterprise", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "description": "Professional cinematic video with native audio (up to 4K)", "supportedModes": [ "generate" ], "order": 5 }, "jobType": "video_generation" }, "video-seedance25": { "active": true, "displayName": "Seedance 2.5", "description": "ByteDance's next-generation Seedance with up to 30-second native clips, up to 50 multimodal references (30 images / 10 videos / 10 audio), and synchronized audio. Supports text, image, first+last frame, and reference-to-video. Currently 480p and 720p.", "dateAdded": "2026-07-07", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "seedance25", "seedance2.5", "seedance-2-5", "seedance-2.5" ], "capabilities": { "video": { "providerModel": "bytedance/seedance-2-5", "providerModelT2V": "bytedance/seedance-2-5", "pricing": { "_perSecond": 300, "_perSecond480p": 150, "_perSecond720p": 300 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 50, "images": 30, "videos": 10, "audio": 10 }, "duration": { "min": 4, "max": 30 }, "resolutions": [ "480p", "720p" ], "aspectRatios": [ "auto", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "acceptsSafetyIdentifier": true, "notes": "Seedance 2.5 via InferenceSH (bytedance/seedance-2-5). Single endpoint handles I2V/T2V/R2V via input fields (image/end_image for first+last frame; reference_images/reference_videos/reference_audios for multimodal); up to 30 reference images, 10 reference videos, 10 reference audios (50 total); duration 4–30s; 480p/720p for now; per-second pricing (150 credits/s @480p, 300/s @720p); audio always included; forwards safety_identifier when an org context is present (optional, not enforced)." } }, "frontend": { "value": "seedance25", "label": "Seedance 2.5", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "description": "Cinematic video up to 30s with up to 50 multimodal references (480p/720p)", "supportedModes": [ "generate" ], "order": 2 }, "jobType": "video_generation" }, "video-seedance25-enterprise": { "active": true, "displayName": "Seedance 2.5 Enterprise", "description": "ByteDance's professional Seedance 2.5 Studio variant with private asset library support. Automatically uploads reference images to BytePlus virtual portrait library for enhanced character consistency. Up to 30s clips with up to 50 multimodal references (30 images / 10 videos / 10 audio) and synchronized audio. Currently 480p and 720p.", "dateAdded": "2026-08-06", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "seedance25-enterprise", "seedance2.5-enterprise", "seedance-2-5-enterprise", "seedance-2.5-enterprise", "seedance-2-5-studio", "seedance25-studio" ], "enterpriseOnly": true, "capabilities": { "video": { "providerModel": "bytedance/seedance-2-5-studio", "providerModelT2V": "bytedance/seedance-2-5-studio", "pricing": { "_perSecond": 300, "_perSecond480p": 150, "_perSecond720p": 300 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "allowed_references": { "total": 50, "images": 30, "videos": 10, "audio": 10 }, "duration": { "min": 4, "max": 30 }, "resolutions": [ "480p", "720p" ], "aspectRatios": [ "auto", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16" ], "features": { "multiFrame": true }, "requireSafetyIdentifier": true, "acceptsSafetyIdentifier": true, "notes": "Seedance 2.5 Enterprise via InferenceSH (bytedance/seedance-2-5-studio); single endpoint handles I2V/T2V/R2V via input fields; private asset library / virtual portrait namespacing via safety_identifier; up to 30 reference images, 10 reference videos, 10 reference audios (50 total); duration 4–30s; 480p/720p for now; per-second pricing (150 credits/s @480p, 300/s @720p); audio always included; enterprise only. Requires a valid LetzAI organizationId (validated against /internal/organizations/{id} and forwarded as InferenceSH safety_identifier)." } }, "frontend": { "value": "seedance25-enterprise", "label": "Seedance 2.5 Enterprise", "providerIcon": "/providers/bytedance.png", "providerName": "ByteDance", "description": "Professional Seedance 2.5 with private asset library and up to 50 refs (480p/720p)", "supportedModes": [ "generate" ], "order": 3 }, "jobType": "video_generation" }, "video-gemini-omni": { "active": true, "displayName": "Gemini Omni Flash", "description": "High-speed multimodal video with native audio — the cheaper, faster Omni-reference alternative.", "dateAdded": "2026-06-30", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "omni", "gemini-omni", "gemini-omni-flash", "gemini-omni-flash-preview", "omni-flash" ], "capabilities": { "video": { "providerModel": "google/gemini-omni-flash", "providerModelT2V": "google/gemini-omni-flash", "submitMode": "gemini-omni-flash", "pricing": { "_perSecond": 100, "_perSecond720p": 100, "_audioMultiplier": 1, "_hqMultiplier": 1 }, "allowAudio": true, "alwaysAudio": true, "supportsReference": true, "supportsR2V": true, "duration": { "min": 3, "max": 10, "step": 1 }, "resolutions": [ "720p", "1080p" ], "aspectRatios": [ "16:9", "9:16" ], "features": { "multiFrame": false }, "notes": "Gemini Omni Flash via InferenceSH (google/gemini-omni-flash), submitted as mode 'gemini-omni-flash'. Schema: prompt (required), image (optional I2V), reference_images (optional), aspect_ratio (16:9|9:16). Audio always included. 100 credits/s; 3-10s (worker clamps for billing/UI; duration is prompt-guided, not an API field); 720p/1080p for pricing; invalid aspect → 16:9. Video/audio reference inputs are not supported on this InferenceSH app — use Seedance 2.0 for full multimodal R2V. withSound always true (hide audio toggle)." } }, "frontend": { "value": "gemini-omni-flash", "label": "Gemini Omni Flash", "providerIcon": "/providers/google.webp", "providerName": "Google", "description": "Fast, affordable Omni-reference video with native audio", "supportedModes": [ "generate" ], "order": 6 }, "jobType": "video_generation" }, "video-grok": { "active": false, "displayName": "Grok Video", "description": "xAI's video generation model with text-to-video and image-to-video support, up to 15 seconds at 720p.", "dateAdded": "2026-03-12", "provider": "xai", "allowBYOK": false, "aliases": [ "grok", "grok-video", "xai-video", "grok-imagine-video" ], "capabilities": { "video": { "providerModel": "grok-imagine-video", "providerModelT2V": "grok-imagine-video", "pricing": { "_perSecond": 40, "_hqMultiplier": 2 }, "allowAudio": false, "supportsReference": true, "duration": { "min": 1, "max": 15 }, "nsfwCheck": true, "resolutions": [ "480p", "720p" ], "aspectRatios": [ "16:9", "9:16", "1:1", "4:3", "3:4", "3:2", "2:3" ], "notes": "xAI Grok Imagine Video via REST API; $0.050/second; async with polling; t2v and i2v" } }, "frontend": { "value": "grok-video", "label": "Grok Video", "providerIcon": "/providers/xai.png", "providerName": "xAI", "description": "Video generation by xAI", "supportedModes": [ "generate" ], "order": 5 }, "jobType": "video_generation" }, "_comment_upscale": "=== UPSCALING MODELS ===", "upscale-letzai": { "active": false, "displayName": "LetzAI Upscaler", "description": "LetzAI's in-house upscaler with three modes: Simple (high speed), Sharp (high quality), and Soft (high consistency).", "dateAdded": "2024-06-01", "provider": "comfyui", "allowBYOK": false, "aliases": [ "simple", "sharp", "soft", "default" ], "capabilities": { "upscaling": { "pricing": { "_flat": 20 }, "notes": "Local ComfyUI upscaler; modes: simple (SeedVR2), sharp (FLUX+UltraSharp), soft (FLUX+NMKD-Siax)" } }, "jobType": "upscale" }, "upscale-nano-banana-pro": { "active": true, "displayName": "Nano Banana Pro Upscaler", "description": "AI-powered upscaling using Google Gemini 3 Pro with intelligent detail enhancement up to 4K resolution.", "dateAdded": "2025-11-20", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "nano-banana-pro-upscale" ], "capabilities": { "upscaling": { "pricing": { "1k": 80, "2k": 160, "4k": 240 }, "defaultMode": "2k", "notes": "Gemini 3 Pro via InferenceSH; mode value 'nano-banana-pro'; size 4→1K, 8→2K, 12→4K" } }, "frontend": { "value": "nano-banana-pro", "label": "Nano Banana Pro", "subLabel": "Up to 4K resolution", "providerIcon": "/providers/google.webp" }, "jobType": "upscale" }, "upscale-nano-banana-2": { "active": true, "displayName": "Nano Banana 2 Upscaler", "description": "Faster AI upscaling using Google Gemini 3.1 Flash with intelligent detail enhancement up to 4K resolution.", "dateAdded": "2026-06-03", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "nano-banana-2-upscale" ], "capabilities": { "upscaling": { "pricing": { "1k": 50, "2k": 100, "4k": 150 }, "defaultMode": "2k", "notes": "Gemini 3.1 Flash via InferenceSH; mode value 'nano-banana-2'; size 4→1K, 8→2K, 12→4K" } }, "frontend": { "value": "nano-banana-2", "label": "Nano Banana 2", "subLabel": "Up to 4K — faster & cheaper", "providerIcon": "/providers/google.webp" }, "jobType": "upscale" }, "upscale-gpt-image-2": { "active": true, "displayName": "GPT Image 2 Upscaler", "description": "AI upscaling via OpenAI GPT Image 2 edit with intelligent detail enhancement up to 4K resolution.", "dateAdded": "2026-07-05", "provider": "fal", "allowBYOK": false, "aliases": [ "gpt-image-2-upscale" ], "capabilities": { "upscaling": { "pricing": { "1k": 160, "2k": 240, "4k": 480 }, "defaultMode": "2k", "notes": "FAL openai/gpt-image-2/edit; mode value 'gpt-image-2'; size 4→1K, 8→2K/HD, 12→4K" } }, "frontend": { "value": "gpt-image-2", "label": "GPT Image 2", "subLabel": "Up to 4K resolution", "providerIcon": "/providers/chatgpt.png" }, "jobType": "upscale" }, "upscale-pruna": { "active": true, "displayName": "Pruna Upscaler", "description": "Fast AI upscaling via Pruna P-Image-Upscale with optional detail and realism enhancement up to 128 megapixels.", "dateAdded": "2026-05-21", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "pruna", "p-image-upscale" ], "capabilities": { "upscaling": { "pricing": { "1mp": 5, "2mp": 10, "4mp": 20, "8mp": 40, "16mp": 80, "32mp": 120, "64mp": 240, "128mp": 400 }, "defaultMode": "4mp", "notes": "Pruna P-Image-Upscale via InferenceSH; mode value 'pruna'; size 4→1MP, 8→4MP, 12→16MP, 14→32MP, 16→64MP, 18→128MP (max 128MP)" } }, "frontend": { "value": "pruna", "label": "Pruna", "subLabel": "Fast upscaling up to 128 MP", "providerIcon": "/providers/pruna.png" }, "jobType": "upscale" }, "upscale-magnific-creative": { "active": false, "displayName": "Magnific Creative", "description": "Freepik's creative upscaler that can add and infer new details guided by text prompts, with up to 16x scaling.", "dateAdded": "2024-12-01", "provider": "freepik", "allowBYOK": false, "aliases": [ "freepik", "magnific-creative" ], "capabilities": { "upscaling": { "pricing": { "2x": 120, "4x": 240, "8x": 960, "16x": 960 }, "defaultMode": "4x", "notes": "Magnific Creative via Freepik API; creative upscale that can add/infer new details; supports prompt" } }, "frontend": { "value": "freepik", "label": "Magnific Creative", "subLabel": "Can add and infer new details", "providerIcon": "/providers/magnific.jpg" }, "jobType": "upscale" }, "upscale-magnific-precision": { "active": false, "displayName": "Magnific Precision V2", "description": "Freepik's faithful upscaler with no hallucinations — preserves text, UI elements, and original image structure with ultimate fidelity.", "dateAdded": "2025-11-10", "provider": "freepik", "allowBYOK": false, "aliases": [ "freepik-precision", "magnific-precision" ], "capabilities": { "upscaling": { "pricing": { "2x": 120, "4x": 240, "8x": 960, "16x": 960 }, "defaultMode": "4x", "notes": "Magnific Precision V2 via Freepik API; faithful upscale — no hallucinations, preserves text/UI; scale 2-16" } }, "frontend": { "value": "freepik-precision", "label": "Magnific Precision V2", "subLabel": "Faithful — preserves text & UI", "providerIcon": "/providers/magnific.jpg" }, "jobType": "upscale" }, "upscale-seedance2": { "active": true, "displayName": "Seedance 2.0 Video Upscale", "description": "Upscale a video to 1080p or 4K via Seedance 2.0 reference-to-video. Max 15s.", "dateAdded": "2026-08-08", "provider": "inferencesh", "allowBYOK": false, "aliases": [ "seedance2-upscale", "seedance-2-upscale" ], "capabilities": { "upscaling": { "mediaType": "video", "providerModel": "bytedance/seedance-2-0", "pricing": { "_perSecond": 500, "_perSecond1080p": 500, "_perSecond4k": 1000 }, "defaultMode": "1080p", "resolutions": [ "1080p", "4k" ], "duration": { "min": 4, "max": 15 }, "notes": "Video upscale via InferenceSH Seedance 2.0 R2V. size 8→1080p, 12→4K. generationSettings.duration (seconds) drives per-second credits. Frontend mode value: seedance2." } }, "frontend": { "value": "seedance2", "label": "Seedance 2.0", "providerIcon": "/providers/bytedance.png" }, "jobType": "upscale" }, "upscale-seedance2-enterprise": { "active": true, "displayName": "Seedance 2.0 Enterprise Video Upscale", "description": "Org-scoped Seedance 2.0 Studio video upscale to 1080p or 4K. Max 15s.", "dateAdded": "2026-08-08", "provider": "inferencesh", "allowBYOK": false, "enterpriseOnly": true, "aliases": [ "seedance2-enterprise-upscale", "seedance-2-enterprise-upscale" ], "capabilities": { "upscaling": { "mediaType": "video", "providerModel": "bytedance/seedance-2-0-studio", "pricing": { "_perSecond": 500, "_perSecond1080p": 500, "_perSecond4k": 1000 }, "defaultMode": "1080p", "resolutions": [ "1080p", "4k" ], "duration": { "min": 4, "max": 15 }, "requireSafetyIdentifier": true, "notes": "Enterprise Seedance Studio R2V. Forwards organizationId as safety_identifier. Frontend mode value: seedance2-enterprise." } }, "frontend": { "value": "seedance2-enterprise", "label": "Seedance 2.0 Enterprise", "providerIcon": "/providers/bytedance.png" }, "jobType": "upscale" }, "upscale-byteplus-video-enhancer": { "active": true, "displayName": "BytePlus Video Enhancer", "description": "Cinema-grade BytePlus MediaKit enhancement (Professional): 30+ algorithms, AI super-res, defect repair, and color. Keeps the original footage (does not regenerate).", "dateAdded": "2026-08-14", "provider": "byteplus", "allowBYOK": false, "aliases": [ "byteplus-video-enhancer", "byteplus-enhancer", "video-enhancer", "mediakit-enhance" ], "capabilities": { "upscaling": { "mediaType": "video", "providerModel": "enhance-video", "pricing": { "_perSecond": 56, "_perSecond1080p": 56, "_perSecond2k": 112, "_perSecond4k": 224, "_perSecond8k": 896, "_fps60Multiplier": 2, "_fps120Multiplier": 4 }, "defaultMode": "1080p", "resolutions": [ "1080p", "2k", "4k", "8k" ], "sizeToResolution": { "8": "1080p", "10": "2k", "12": "4k", "16": "8k" }, "fps": { "default": null, "options": [ 60, 120 ], "min": 15, "max": 120 }, "duration": { "min": 1, "max": 600 }, "notes": "BytePlus AI MediaKit POST /api/v1/tools/enhance-video. Credits = BytePlus VOD Pro ≤30fps USD/min (1080p $4.13, 2K $8.26, 4K $16.53, 8K $66.12) at the MiniMax house rate ($0.26/s → 210 credits/s ≈ 808 credits/$). fps 60 = 2×, 120 = 4× (same as BytePlus fps bands). Defaults: professional, hd, high bitrate, source fps. size 8→1080p, 10→2K, 12→4K, 16→8K." } }, "frontend": { "value": "byteplus-video-enhancer", "label": "BytePlus Video Enhancer", "subLabel": "Professional cinema enhance — 1080p to 8K", "providerIcon": "/providers/bytedance.png" }, "jobType": "upscale" }, "_comment_deprecated": "=== DEPRECATED MODELS ===", "letzai-v1": { "active": false, "displayName": "LetzAI V1", "description": "First generation LetzAI image model built on Stable Diffusion 1.5 with custom fine-tuning for creative generation.", "dateAdded": "2023-06-01", "provider": "comfyui", "allowBYOK": false, "aliases": [ "letzai-v1", "letzaiv1" ], "capabilities": { "generation": { "pricing": { "_flat": 5 }, "notes": "Deprecated - replaced by V2" } }, "deprecated": true }, "letzai-v2": { "active": false, "displayName": "LetzAI V2", "description": "Second generation LetzAI image model built on SDXL with improved detail, resolution, and prompt understanding.", "dateAdded": "2024-01-15", "provider": "comfyui", "allowBYOK": false, "aliases": [ "letzai-v2", "letzaiv2" ], "capabilities": { "generation": { "pricing": { "_flat": 10 }, "notes": "Deprecated - replaced by V3" } }, "deprecated": true }, "letzai-v3": { "active": false, "displayName": "LetzAI V3", "description": "Third generation LetzAI image model with enhanced quality and photorealism based on Flux1.dev.", "dateAdded": "2024-09-01", "provider": "comfyui", "allowBYOK": false, "aliases": [ "letzai-v3", "letzaiv3" ], "capabilities": { "generation": { "pricing": { "_flat": 15 }, "notes": "Deprecated - replaced by third-party models" } }, "deprecated": true } }, "providers": { "anthropic": { "displayName": "Anthropic", "envKeys": [ "ANTHROPIC_API_KEY" ], "notes": "Claude text models (Sonnet 4.6/5, Opus 5/4.8, Fable 5) for the chat + canvas agent harness" }, "vertex": { "displayName": "Google Vertex AI", "envKeys": [ "GOOGLE_CLOUD_PROJECT" ], "notes": "Gemini models; can be toggled to InferenceSH via USEVERTEX=FALSE" }, "inferencesh": { "displayName": "InferenceSH", "envKeys": [ "INF_API_TOKEN" ], "supportsOrgKey": true, "notes": "Gemini 3 Pro, Pruna P-Image-Upscale; supports per-org API key override via orgId" }, "fal": { "displayName": "FAL.ai", "envKeys": [ "FAL_KEY" ], "icon": "/providers/fal.svg", "notes": "Flux family, Kling, Nano Banana" }, "byteplus": { "displayName": "BytePlus", "envKeys": [ "ARK_API_KEY", "BYTEPLUS_MEDIAKIT_API_KEY" ], "notes": "Seedream models; Seedance 2.0 standard when USE_BYTEPLUS_SEEDANCE=TRUE (InferenceSH when FALSE). Video enhancement uses BYTEPLUS_MEDIAKIT_API_KEY (AI MediaKit, not ARK)." }, "xai": { "displayName": "xAI (Grok)", "envKeys": [ "XAI_API_KEY" ], "notes": "Grok 4.6 text (chat + canvas agent) and Grok Imagine Image 2.0 / Image Quality / video via the xAI API" }, "openai": { "displayName": "OpenAI", "envKeys": [ "OPENAI_API_KEY" ], "notes": "GPT-5.5 and GPT-5.6 Sol text models for the chat + canvas agent harness." }, "freepik": { "displayName": "Freepik (Magnific)", "envKeys": [ "FREEPIK_API_KEY" ], "notes": "Magnific Creative and Precision V2 upscalers" }, "comfyui": { "displayName": "Local", "envKeys": [], "notes": "Local GPU workflows for LetzAI upscaling" } } }