Alibaba Cloud
ASRQwen Audio 3.1 ASR Flash Message
qwen-audio-3.1-asr-flash-message DashScope real-time speech recognition for short utterances such as voice messages and IME input. Billed by input and output tokens. Context 8K, max output 1K.
Model Catalog
Compare model capabilities, context, modalities, and catalog pricing.
A starting point, not a limitation
Learn about custom configuration This catalog lists the static presets currently bundled with OctaFuse Gateway for direct Admin import. Gateway still lets you create any custom Provider, Model, or compatible endpoint. Not listed does not mean unsupported.
Alibaba Cloud
ASRQwen Audio 3.1 ASR Flash Streaming
qwen-audio-3.1-asr-flash-streaming DashScope real-time speech recognition for continuous transcription. Billed by input and output tokens. Context 8K, max output 1K.
Alibaba Cloud
ASRQwen Audio 3.1 ASR Flash Filetrans
qwen-audio-3.1-asr-flash-filetrans DashScope asynchronous speech recognition for recorded files up to 12 hours, with speaker diarization. Billed by input and output tokens. Context 8K, max output 1K.
Alibaba Cloud
ASRQwen Audio 3.1 ASR Flash
qwen-audio-3.1-asr-flash DashScope synchronous speech recognition for recordings up to 5 minutes, with speaker diarization. Billed by input and output tokens. Context 8K, max output 1K.
Alibaba Cloud
ASRQwen Audio 3.0 ASR Flash Streaming
qwen-audio-3.0-asr-flash-streaming DashScope real-time speech recognition for low-latency streaming audio.
Alibaba Cloud
ASRQwen Audio 3.0 ASR Flash Filetrans
qwen-audio-3.0-asr-flash-filetrans DashScope asynchronous speech recognition for recorded audio files.
Alibaba Cloud
ASRQwen Audio 3.0 ASR Flash
qwen-audio-3.0-asr-flash DashScope synchronous speech recognition for short recorded audio.
Alibaba Cloud
ASRQwen3 ASR Flash Realtime
qwen3-asr-flash-realtime DashScope Qwen3 ASR real-time speech recognition over WebSocket sessions.
Alibaba Cloud
ASRQwen3 ASR Flash Filetrans
qwen3-asr-flash-filetrans DashScope Qwen3 ASR asynchronous speech recognition for recorded audio files.
Alibaba Cloud
ASRQwen3 ASR Flash
qwen3-asr-flash DashScope Qwen3 ASR for synchronous audio transcription.
Alibaba Cloud
ASRFun-ASR Realtime
fun-asr-realtime DashScope Fun-ASR real-time speech recognition over WebSocket streaming.
Alibaba Cloud
ASRFun-ASR
fun-asr DashScope Fun-ASR asynchronous speech recognition for recorded audio files.
Alibaba Cloud
ASRQwen Audio 3.1 TTS Flash
qwen-audio-3.1-tts-flash DashScope low-latency text-to-speech with instruction control and voice cloning. Billed by input and output tokens.
Alibaba Cloud
TTSQwen Audio 3.0 TTS Plus
qwen-audio-3.0-tts-plus DashScope high-quality text-to-speech model billed by generated characters.
Alibaba Cloud
TTSQwen Audio 3.0 TTS Flash
qwen-audio-3.0-tts-flash DashScope low-latency text-to-speech model billed by generated characters.
Alibaba Cloud
TTSCosyVoice V1
cosyvoice-v1 DashScope text-to-speech model with streaming synthesis support.
Alibaba Cloud
TTSCosyVoice V2
cosyvoice-v2 DashScope text-to-speech model with streaming synthesis support.
Alibaba Cloud
TTSCosyVoice V3 Flash
cosyvoice-v3-flash DashScope low-latency text-to-speech model billed by generated characters.
Alibaba Cloud
TTSCosyVoice V3 Plus
cosyvoice-v3-plus DashScope high-quality text-to-speech model billed by generated characters.
Alibaba Cloud
TTSCosyVoice V3.5 Flash
cosyvoice-v3.5-flash DashScope CosyVoice 3.5 Flash TTS with voice cloning/design and instruction control; billed by generated characters.
Alibaba Cloud
TTSCosyVoice V3.5 Plus
cosyvoice-v3.5-plus DashScope CosyVoice 3.5 Plus high-expressiveness TTS with voice cloning/design and instruction control; billed by generated characters.
Alibaba Cloud
Qwen3.7 Text Embedding
qwen3.7-text-embedding Alibaba Cloud Model Studio text embedding model. Billed by input tokens only. Use the OpenAI-compatible POST /v1/embeddings.
Alibaba Cloud
Text Embedding V4
text-embedding-v4 Alibaba Cloud Model Studio text embedding V4 (successor of V3). Billed by input tokens only. Use the OpenAI-compatible POST /v1/embeddings.
Alibaba Cloud
ImageQwen Image 3.0 Pro
qwen-image-3.0-pro Alibaba's flagship Qwen Image 3.0 model for dense layouts, multilingual text rendering, and photographic detail. Supports text-to-image and image editing up to 2K.
Alibaba Cloud
ImageQwen Image 3.0
qwen-image-3.0 A faster Qwen Image 3.0 model for general generation and editing. Same 2K ceiling as Pro, with a flat per-image output price.
Alibaba Cloud
ImageQwen Image 2.1 Pro
qwen-image-2.1-pro A unified text-to-image and image editing model with 7B visual generation parameters. Supports up to 10 reference images, transparent RGBA output, local editing, and subject preservation.
Alibaba Cloud
ImageQwen Image 2.1 Turbo
qwen-image-2.1-turbo A cost-effective Qwen-Image 2.1 open-source model for image generation and editing. Text/image input, image output.
Alibaba Cloud
ImageWan 2.7 Image Pro
wan2.7-image-pro Alibaba Wan 2.7 flagship for text-to-image, sequential sets, multi-reference editing, brand color control, and up to 4K generation.
Alibaba Cloud
ImageWan 2.7 Image
wan2.7-image A faster Wan 2.7 model for generation, sequential sets, and multi-reference editing, with a 2K output ceiling.
Alibaba Cloud
LLMQwen3.8 Max
qwen3.8-max A 2.4T-parameter MoE flagship for coding, office productivity, vision, and long-horizon agents. Native multimodal understanding supports long documents and video, with strong end-to-end delivery on professional tasks.
Alibaba Cloud
LLMQwen3.8 Max Prime
qwen3.8-max-prime A high-throughput variant of Qwen3.8 Max optimized for faster inference. Supports both reasoning and non-reasoning modes with the same 1M context window.
Alibaba Cloud
LLMQwen3.8 2.4T-A95B
qwen3.8-2.4t-a95b A 2.4T-parameter open-source flagship for complex reasoning, coding, and knowledge work. Supports both reasoning and non-reasoning modes with a 1M context window.
Alibaba Cloud
LLMQwen3.8 27B
qwen3.8-27b A 27B-parameter multimodal Qwen model for efficient reasoning, coding, and visual understanding. Supports both reasoning and non-reasoning modes with image and video input.
Alibaba Cloud
LLMQwen3.8 Omni Flash
qwen3.8-omni-flash A multimodal Qwen model for text, image, audio, and video understanding. Supports deep thinking with adjustable reasoning intensity for complex tasks.
Alibaba Cloud
LLMQwen3.8 Max Preview
qwen3.8-max-preview A flagship Qwen preview for complex reasoning, coding, vision, and long-horizon agentic workloads. Preview channel for early access; production traffic should prefer qwen3.8-max.
Alibaba Cloud
LLMQwen3.8 Flash
qwen3.8-flash A fast multimodal Qwen3.8 model with a native 1M context window for coding, agents, documents, and visual understanding. Near-flagship quality at a fraction of Qwen3.8 Max cost, with image and video input.
Alibaba Cloud
LLMQwen3.7 Plus
qwen3.7-plus A cost-efficient multimodal Qwen model for general reasoning, vision, and production applications. It combines text and image understanding with a cost-conscious design for production-scale multimodal reasoning.
Alibaba Cloud
LLMQwen3.7 Max
qwen3.7-max The Qwen3.7 flagship, optimized for coding, productivity, and agent-centric workloads. Its agent-oriented design emphasizes coding, office productivity, and complex multi-step work.
Alibaba Cloud
LLMQwen3.7 Flash
qwen3.7-flash A fast multimodal Qwen3.7 model with stronger vision, agent execution, and coding throughput than Qwen3.6 Flash. Built for low-latency multimodal understanding and agent workflows.
Alibaba Cloud
LLMQwen3.6 Plus
qwen3.6-plus A scalable long-context Qwen model balancing strong reasoning with efficient inference. A hybrid of efficient linear attention and sparse expert routing improves scalability and inference efficiency.
Alibaba Cloud
LLMQwen3.6 Max
qwen3.6-max A high-capability Qwen model for demanding reasoning, coding, and knowledge work.
Alibaba Cloud
LLMQwen3.6 flash
qwen3.6-flash A fast multimodal Qwen model with a large context window for high-throughput workloads. It handles text, image, and video inputs with a long context window for latency-sensitive multimodal applications.
Alibaba Cloud
LLMQwen3.5 Plus
qwen3.5-plus A balanced Qwen model for general chat, reasoning, coding, and multimodal tasks.
Alibaba Cloud
LLMQwen3.5 Flash
qwen3.5-flash A low-latency Qwen model for responsive chat, extraction, and lightweight agents.
Alibaba Cloud
LLMQwen3-VL Plus
qwen3-vl-plus A vision-language model for document, image, screen, and visual reasoning tasks.
Alibaba Cloud
LLMQwen3.5 Omni Plus
qwen3.5-omni-plus A full-capability Qwen model for understanding text, images, audio, and video.
Alibaba Cloud
LLMQwen3.5 Omni Flash
qwen3.5-omni-flash A fast omni-modal Qwen model for real-time multimodal understanding and interaction.
Anthropic
LLMClaude Sonnet 5.5
claude-sonnet-5-5 Anthropic's faster Sonnet for coding, agents, and professional work. Adaptive thinking is on by default and cannot be fully disabled; the default effort is high, and between_tools turns off up-front thinking at high effort or below. Forced tool_choice values any and tool return an error, as do non-default temperature, top_p, and top_k.
Anthropic
LLMClaude Haiku 5.5
claude-haiku-5-5 Anthropic's fastest Claude 5 model for latency-sensitive tasks. Adaptive thinking is on by default with medium effort and supports low through max; thinking can be disabled at high effort or below. Prompts over 100K tokens are billed at a higher tier. Manual budget_tokens returns 400, as do non-default temperature, top_p, and top_k.
Anthropic
LLMClaude Opus 5.5
claude-opus-5-5 Anthropic's model for long-running agentic coding and knowledge work. Adaptive thinking is always on and cannot be disabled; the default effort is medium. Forced tool_choice values any and tool return an error.
Anthropic
LLMClaude Fable 5.1
claude-fable-5-1 Anthropic's successor to Claude Fable 5 for long-running agentic coding, multistep research, and document work. Adaptive thinking is always on; input and output prices match Fable 5, with cheaper prompt-cache reads.
Anthropic
LLMClaude Fable 5
claude-fable-5 An Anthropic model built for autonomous knowledge work, coding, and multimodal agents. It is positioned for autonomous knowledge work and coding with multimodal input, reasoning, and tool-oriented workflows.
Anthropic
LLMClaude Sonnet 5
claude-sonnet-5 A frontier Sonnet model for coding, professional work, and adaptive agentic reasoning. Adaptive reasoning levels let applications balance frontier coding and agent performance against latency and token use.
Anthropic
LLMClaude Opus 5
claude-opus-5 Anthropic's flagship for demanding reasoning, coding, visual analysis, and long-running agents. It is especially capable at end-to-end software tasks, code review and bug finding, document and chart analysis, complex office deliverables, tool use, and coordinating parallel subagents.
Anthropic
LLMClaude Opus 4.8
claude-opus-4.8 A high-capability Opus model for long-context reasoning, coding, and multimodal analysis. It combines multimodal input, reasoning support, and a long context window for demanding general-purpose workflows.
Anthropic
LLMClaude Opus 4.7
claude-opus-4.7 An Opus model designed for long-running asynchronous agents and complex software workflows. The model is designed for asynchronous agents that sustain coding and professional work over extended runs.
Anthropic
LLMClaude Opus 4.6
claude-opus-4.6 A powerful Claude model for end-to-end coding and long-running professional workflows. It is optimized for agents that operate across complete software and professional workflows instead of isolated prompts.
Anthropic
LLMClaude Sonnet 4.6
claude-sonnet-4.6 A balanced frontier model for iterative development, agents, and professional knowledge work. Strengths include iterative development, complex codebase navigation, and end-to-end project execution with tools.
Anthropic
LLMClaude Haiku 4.5
claude-haiku-4.5 A fast, efficient Claude model for latency-sensitive chat, coding, and agent tasks. It targets near-frontier intelligence at substantially lower latency and cost for high-volume interactive workloads.
ByteDance
ImageDoubao Seedream 5.0 Flash
doubao-seedream-5-0-flash A fast, cost-efficient Seedream model for high-volume image generation, editing, transparent backgrounds, and layer decomposition.
ByteDance
ImageDoubao Seedream 5.0 Pro
doubao-seedream-5-0-pro A premium Seedream model for high-fidelity image generation and professional visual creation.
ByteDance
ImageDoubao Seedream 5.0
doubao-seedream-5-0 A Seedream model for general image generation, editing, and visual design workflows.
ByteDance
LLMDoubao Seed 2.0 Mini
doubao-seed-2-0-mini-260215 A compact Doubao model for fast, cost-sensitive chat and lightweight reasoning.
ByteDance
LLMDoubao Seed 2.0 Mini (260428)
doubao-seed-2-0-mini-260428 A compact all-modal Doubao model for deep thinking across text, image, video, and audio. Successor to doubao-seed-2-0-mini-260215. Context window is 256K, maximum input is 224K, and maximum answer is 128K. Official audio-input list prices are not stored in this preset: CNY 3 / 6 / 12 per million tokens (cache hit 0.6 / 1.2 / 2.4) for input up to 32K / 128K / 256K; USD 1.50 / 3.00 (cache hit 0.30 / 0.60) for input up to 128K / 256K.
ByteDance
LLMDoubao Seed 2.0 Lite
doubao-seed-2-0-lite-260215 A lightweight Doubao model optimized for high-throughput text and vision workloads.
ByteDance
LLMDoubao Seed 2.0 Lite (260428)
doubao-seed-2-0-lite-260428 A lightweight all-modal Doubao model for deep thinking across text, image, video, and audio. Successor to doubao-seed-2-0-lite-260215. Context window is 256K, maximum input is 224K, and maximum answer is 128K. Official audio-input list prices are not stored in this preset: CNY 9 / 13.5 / 27 per million tokens (cache hit 1.8 / 2.7 / 5.4) for input up to 32K / 128K / 256K; USD 3.75 / 7.50 (cache hit 0.75 / 1.50) for input up to 128K / 256K.
ByteDance
LLMDoubao Seed 2.0 Pro
doubao-seed-2-0-pro-260215 A capable Doubao model for advanced reasoning, multimodal understanding, and agents.
ByteDance
LLMDoubao Seed 2.1 Pro
doubao-seed-2-1-pro-260628 A newer Doubao Pro model for complex reasoning, coding, and multimodal production use.
ByteDance
LLMDoubao Seed 2.1 Pro (260915)
doubao-seed-2-1-pro-260915 A Doubao Seed 2.1 Pro model for deep thinking, tool calling, structured output, GUI tasks, and multimodal understanding of text, images, and video. Context window is 1024K and maximum answer is 256K. BytePlus ModelArk does not list this model, so there is no USD price.
ByteDance
LLMDoubao Seed 2.1 Lite (260915)
doubao-seed-2-1-lite-260915 A Doubao Seed 2.1 Lite model for deep thinking, tool calling, structured output, GUI tasks, and multimodal understanding of text, images, video, and audio. Context window is 1024K and maximum answer is 256K. Official audio-input list price is not stored in this preset: CNY 12 per million tokens (cache hit 2.4). BytePlus ModelArk does not list this model, so there is no USD price.
ByteDance
LLMDoubao Seed 2.1 Turbo
doubao-seed-2-1-turbo-260628 A low-latency Doubao model for responsive multimodal applications and agents.
ByteDance
LLMDoubao Seed Evolving
doubao-seed-evolving An evolving Doubao model intended for adaptive reasoning and agentic experimentation.
DeepSeek
LLMDeepSeek V4.1 Flash
deepseek-v4.1-flash The official DeepSeek V4.1 Flash model for fast long-context reasoning, coding, and agents. In internal and external evaluations, it comprehensively surpasses V4 Pro in performance, cost, speed, and total time.
DeepSeek
LLMDeepSeek V4 Pro
deepseek-v4-pro The official DeepSeek V4 Pro flagship for advanced reasoning, coding, and sustained agent workloads. The 0813 production checkpoint keeps 1M context and dual thinking modes under the same API name.
DeepSeek
LLMDeepSeek V4 Flash
deepseek-v4-flash The official DeepSeek V4 Flash model for fast long-context reasoning, coding, and agents. The 0731 post-training refresh keeps the mixture-of-experts architecture while strengthening agentic coding and tool use.
DeepSeek
LLMDeepSeek-V3.2
deepseek-v3.2 An efficient DeepSeek model combining strong reasoning with agentic tool use. Its sparse-attention design targets efficient long-context reasoning while preserving strong tool-use performance.
ElevenLabs
ASRElevenLabs Scribe v2
elevenlabs-scribe-v2 ElevenLabs Scribe v2 speech-to-text for POST /v1/speech-to-text (file transcription): 90+ languages, word-level timestamps, speaker diarization, and audio event tagging. Set the route provider model name to scribe_v2. Use elevenlabs-asr-file for OpenAI /v1/audio/transcriptions, or ElevenLabs audio.transcriptions passthrough for the native endpoint. Official list price $0.22 per hour; gateway meters by audio duration (per second). Optional keyterm prompting ($0.05/hour) and transcript editing ($0.07/hour) are billed extra by ElevenLabs and not included. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
ASRElevenLabs Scribe v2 Medical
elevenlabs-scribe-v2-medical ElevenLabs Scribe v2 Medical speech-to-text, a Scribe v2 fine-tune for clinical audio with better recognition of drug names, anatomy, and pathology terms. Same API and price as Scribe v2: POST /v1/speech-to-text, 90+ languages, word-level timestamps, speaker diarization. Set the route provider model name to scribe_v2_medical. Use elevenlabs-asr-file for OpenAI /v1/audio/transcriptions, or ElevenLabs audio.transcriptions passthrough for the native endpoint. Official list price $0.22 per hour; gateway meters by audio duration (per second). Batch only; optional keyterm prompting and transcript editing are billed extra by ElevenLabs and not included. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
ASRElevenLabs Scribe v2 Realtime
elevenlabs-scribe-v2-realtime ElevenLabs Scribe v2 Realtime speech-to-text over WebSocket (GET /v1/speech-to-text/realtime): low-latency streaming transcription with partial and committed transcripts, 90+ languages. Set the route provider model name to scribe_v2_realtime and use the ElevenLabs audio.transcriptions.realtime passthrough; the OpenAI /v1/audio/transcriptions endpoint is not supported for this model. Official list price $0.39 per hour; the gateway meters per second from the audio bytes the client sends (using the session audio_format), so it is an estimate to reconcile against the ElevenLabs dashboard. Browsers cannot set the xi-api-key header: authenticate with the Sec-WebSocket-Protocol token octafuse-api-key.<key>. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Eleven v4
elevenlabs-v4 ElevenLabs Eleven v4 text-to-speech: the most emotive, highest-quality model with strong voice cloning, 90+ languages, multi-speaker dialogue, and a 10,000-character limit per request. Set the route provider model name to eleven_v4. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.08 per 1K characters; billed per character. The gateway proxies the HTTP text-to-speech endpoints and the single-stream WebSocket stream-input endpoint (multi-stream-input is not supported); the WebSocket is billed by the characters the client sends. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Eleven v4 Turbo
elevenlabs-v4-turbo ElevenLabs Eleven v4 Turbo text-to-speech: real-time model with median inference latency of about 100 ms, v4-style expressiveness, audio tags, and 90+ languages. Set the route provider model name to eleven_v4_turbo. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.04 per 1K characters; billed per character. The gateway proxies the HTTP text-to-speech endpoints and the single-stream WebSocket stream-input endpoint (multi-stream-input is not supported); the WebSocket is billed by the characters the client sends. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Eleven v3
elevenlabs-v3 ElevenLabs Eleven v3 text-to-speech: emotionally rich, expressive delivery with audio tags and multi-speaker dialogue, 70+ languages. Set the route provider model name to eleven_v3. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.08 per 1K characters; billed per character. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Eleven v3 Conversational
elevenlabs-v3-conversational ElevenLabs Eleven v3 Conversational text-to-speech: low-latency (about 280 ms) v3 for realtime conversation, with audio tags and 70+ languages. Set the route provider model name to eleven_v3_conversational. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.04 per 1K characters; billed per character. The gateway proxies the HTTP text-to-speech endpoints and the single-stream WebSocket stream-input endpoint (multi-stream-input is not supported); the WebSocket is billed by the characters the client sends. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Multilingual v2
elevenlabs-multilingual-v2 ElevenLabs Multilingual v2 text-to-speech: lifelike, consistent quality for long-form audio, 29 languages. Set the route provider model name to eleven_multilingual_v2. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.08 per 1K characters; billed per character. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Flash v2.5
elevenlabs-flash-v2.5 ElevenLabs Flash v2.5 text-to-speech: ultra-low latency (~75 ms), fast and affordable, 32 languages. Set the route provider model name to eleven_flash_v2_5. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.04 per 1K characters; billed per character. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Flash v2
elevenlabs-flash-v2 ElevenLabs Flash v2 text-to-speech: ultra-low latency (about 75 ms), English only, 30,000-character limit per request. Set the route provider model name to eleven_flash_v2. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.04 per 1K characters; billed per character. The gateway proxies the HTTP text-to-speech endpoints and the single-stream WebSocket stream-input endpoint (multi-stream-input is not supported); the WebSocket is billed by the characters the client sends. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
ElevenLabs
TTSElevenLabs Turbo v2.5
elevenlabs-turbo-v2.5 ElevenLabs Turbo v2.5 text-to-speech: first-generation low-latency model, 32 languages. ElevenLabs marks it deprecated and recommends Flash v2.5, which has lower latency. Set the route provider model name to eleven_turbo_v2_5. Use the elevenlabs-tts adapter for OpenAI /v1/audio/speech (voice is an ElevenLabs voice_id), or ElevenLabs audio.speech passthrough for the native text-to-speech endpoints. Official list price $0.04 per 1K characters; billed per character. The gateway proxies the HTTP text-to-speech endpoints and the single-stream WebSocket stream-input endpoint (multi-stream-input is not supported); the WebSocket is billed by the characters the client sends. CNY price is the USD list price × 7 (placeholder, not an official RMB list price).
Gemini Embedding 2
gemini-embedding-2 Google's multimodal embedding model that maps text, images, video, audio, and PDFs into one embedding space. The catalog uses the Standard text-input price; image, audio, and video inputs have separate list prices and are not modeled here. Use :embedContent or :batchEmbedContents.
Gemini Nano Banana 2.1
gemini-nano-banana-2.1 An update to Nano Banana 2, built for high-efficiency image generation and conversational editing with improved visual quality, multi-turn character consistency, accurate text rendering, and search-grounded generation across 1K, 2K, and 4K resolutions.
Gemini 3.1 Flash Image
gemini-3.1-flash-image A fast Google model for high-quality image generation and editing with multimodal reasoning. Also known as Nano Banana 2, it combines fast generation with advanced editing and multimodal visual reasoning.
Gemini 3 Pro Image Preview
gemini-3-pro-image-preview Google's advanced image generation preview for grounded creation and detailed visual editing. Built on Gemini 3 Pro, it emphasizes grounded creation, stronger multimodal reasoning, and detailed image editing.
Gemini 4 Argon
gemini-4-argon Google's flagship Gemini reasoning model for complex software engineering, scientific problem solving, and multi-step agents. Input and output share a ~1M window; synchronous output defaults to 64K, and long decode can reach ~1M.
Gemini 3.8 Flash
gemini-3.8-flash Our most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows. It improves software engineering, agentic tasks, and multi-step reasoning over Gemini 3.7 Flash.
Gemini 3.7 Flash
gemini-3.7-flash A high-efficiency Gemini workhorse for coding, agents, and knowledge work. It improves first-pass code accuracy, multi-step tool use, and production-ready application output over Gemini 3.6 Flash.
Gemini 3.6 Flash
gemini-3.6-flash A high-efficiency Gemini model for coding, agents, and web or application development. It focuses on polished coding and application outputs with fewer unnecessary edits across agentic workflows.
Gemini 3.1 Pro Preview
gemini-3.1-pro-preview A frontier Gemini reasoning preview for software engineering and reliable agentic workflows. The preview improves software-engineering performance, agent reliability, and token efficiency across complex workflows.
Gemini 3 Flash Preview
gemini-3-flash-preview A fast Gemini thinking model for multi-turn chat, coding, and agent workflows. It offers near-Pro reasoning and tool use at Flash-class speed for multi-turn chat, coding, and agents.
Gemini 3.5 Flash
gemini-3.5-flash An efficient multimodal Gemini model with strong coding, reasoning, and parallel-agent performance. It brings near-Pro coding and reasoning to a faster tier and is tuned for parallel agent execution.
Gemini 3.5 Flash Lite
gemini-3.5-flash-lite A lightweight Gemini model for focused subagents and high-volume production tasks. The lightweight design is especially suited to focused subagents inside larger multi-agent systems.
Gemini 3.1 Flash Lite
gemini-3.1-flash-lite A high-efficiency Gemini preview optimized for high-volume, latency-sensitive applications. It is tuned for high-volume workloads while approaching the quality of larger Flash models.
Gemini 2.5 Pro
gemini-2.5-pro A capable Gemini reasoning model for coding, mathematics, science, and multimodal analysis. Built-in thinking supports more deliberate answers across advanced coding, mathematics, science, and multimodal analysis.
Gemini 2.5 Flash
gemini-2.5-flash A versatile Gemini workhorse balancing reasoning quality, speed, and multimodal support. It balances built-in reasoning with workhorse throughput for coding, mathematics, science, and multimodal production tasks.
Gemini 2.5 Flash Lite
gemini-2.5-flash-lite A low-cost Gemini reasoning model optimized for throughput and fast token generation. The model prioritizes ultra-low latency, low cost, high throughput, and fast token generation.
Meituan LongCat
LLMLongCat 2.0
longcat-2.0 A sparse expert model from Meituan for coding, repository work, and long-horizon agents. Its sparse expert architecture is built for repository-scale coding, long-horizon problem solving, and agentic execution.
MiniMax
ASRMiniMax ASR 1.0
minimax-asr-1.0 MiniMax file speech recognition (asr-1.0) via POST /v1/speech_to_text. Set the route provider model name to asr-1.0 and the adapter to minimax-asr-file. Billed per second of upstream duration, with no minimum.
MiniMax
TTSMiniMax Speech 2.8 HD
minimax-speech-2.8-hd MiniMax sync speech synthesis (speech-2.8-hd) via POST /v1/t2a_v2. Set the route provider model name to speech-2.8-hd. Use minimax-tts for OpenAI /v1/audio/speech, or MiniMax audio.speech passthrough for the native endpoint. Billed per upstream usage_characters.
MiniMax
TTSMiniMax Speech 2.8 Turbo
minimax-speech-2.8-turbo MiniMax low-latency sync speech synthesis (speech-2.8-turbo) via POST /v1/t2a_v2. Set the route provider model name to speech-2.8-turbo. Use minimax-tts for OpenAI /v1/audio/speech, or MiniMax audio.speech passthrough for the native endpoint. Billed per upstream usage_characters.
MiniMax
ImageMiniMax Image 01
minimax-image-01 MiniMax text-to-image and subject-reference image generation (image-01) via POST /v1/image_generation. Set the route provider model name to image-01. Use minimax-image for OpenAI /v1/images/generations, or MiniMax images.generations passthrough for the native endpoint. Billed per successful image.
MiniMax
ImageMiniMax Image 01 Live
minimax-image-01-live MiniMax image-01-live generation via POST /v1/image_generation, including the live style. Set the route provider model name to image-01-live. Use minimax-image for OpenAI /v1/images/generations, or MiniMax images.generations passthrough for the native endpoint. Billed per successful image.
MiniMax
LLMMiniMax M2.5
minimax-m2.5 A MiniMax model for practical productivity, coding, and real-world digital work. Training across varied digital work environments extends its coding foundation into practical, end-to-end productivity tasks.
MiniMax
LLMMiniMax M2.7
minimax-m2.7 An agentic MiniMax model for autonomous productivity and multi-agent collaboration. It emphasizes autonomous productivity, multi-agent collaboration, and continuous improvement in real-world work.
MiniMax
LLMMiniMax M2.7 Highspeed
minimax-m2.7-highspeed A high-speed variant of MiniMax M2.7 for faster inference at the same capability level. Optimized for latency-sensitive productivity and agent workflows.
MiniMax
LLMMiniMax M3
minimax-m3 A long-context multimodal MiniMax model for coding, knowledge work, and agents. The multimodal foundation is designed for long-horizon agents, coding, and knowledge work over extended context.
MiniMax
LLMMiniMax M3.1 Flash
minimax-m3.1-flash The latest M-series language model for agentic reasoning, tool use, coding, and long-context tasks. A frontier multimodal coding model with a 1M context window and tunable thinking depth: an ultra-efficient open-weight model for everyday work. Input includes text, images, and video.
Moonshot AI
LLMKimi K2.5
kimi-k2.5 A native multimodal Kimi model for visual coding and coordinated agent workflows. It pairs native multimodal understanding with visual coding and a self-directed agent-swarm approach.
Moonshot AI
LLMKimi K2.6
kimi-k2.6 A multimodal Kimi model for long-horizon coding, UI generation, and agent orchestration. It targets long-horizon coding, code-driven UI generation, and coordinated multi-agent execution across complex projects.
Moonshot AI
LLMKimi K2.7 Code
kimi-k2.7-code A coding-focused Kimi model for reliable end-to-end programming across long contexts. The coding-focused design aims to complete end-to-end programming tasks reliably across long contexts.
Moonshot AI
LLMKimi K2.7 Code Highspeed
kimi-k2.7-code-highspeed A high-speed variant of Kimi K2.7 Code for faster inference on coding tasks. Optimized for latency-sensitive coding workflows and agent applications.
Moonshot AI
LLMKimi K3
kimi-k3 An open-weight multimodal Kimi model for complex coding, reasoning, and agentic work. The open-weight multimodal reasoning model is positioned for complex coding, knowledge work, and long-running agents.
Moonshot AI
LLMKimi K3.1 Pro
kimi-k3.1-pro The first K3.1-series flagship and Kimi's most capable 3T-class open-weight model, with native visual understanding and a 1M context. It improves coding, 3D creation, financial research, and long-horizon agent work.
Moonshot AI
LLMKimi K3.1 Pro Highspeed
kimi-k3.1-pro-highspeed The high-speed variant of Kimi K3.1 Pro with the same intelligence and a 1M context, at about 2-3× the speed. It stays in the multimodal model list and keeps the same parameter constraints as K3.1 Pro.
OpenAI
ASRWhisper
whisper-1 OpenAI Whisper (whisper-1) speech-to-text for /v1/audio/transcriptions. Official price $0.006/minute; gateway meters by duration (per second).
OpenAI
ASRGPT-4o mini Transcribe
gpt-4o-mini-transcribe OpenAI GPT-4o mini speech-to-text for /v1/audio/transcriptions (json/text). Official rates $1.25/$5 per 1M audio tokens; gateway meters by tokens.
OpenAI
ASRGPT-4o Transcribe
gpt-4o-transcribe OpenAI GPT-4o speech-to-text for /v1/audio/transcriptions (json/text). Official rates $2.50/$10 per 1M audio tokens; gateway meters by tokens.
OpenAI
ASRGPT-4o Transcribe Diarize
gpt-4o-transcribe-diarize OpenAI GPT-4o transcription with speaker diarization (/v1/audio/transcriptions; json/text/diarized_json). Same token rates as gpt-4o-transcribe ($2.50/$10 per 1M); gateway meters by tokens.
OpenAI
Text Embedding 3 Small
text-embedding-3-small OpenAI's small, efficient text embedding model (1536 dimensions by default; supports the dimensions parameter). Billed by input tokens via POST /v1/embeddings.
OpenAI
Text Embedding 3 Large
text-embedding-3-large OpenAI's most capable text embedding model (3072 dimensions by default; supports the dimensions parameter). Billed by input tokens via POST /v1/embeddings.
OpenAI
Text Embedding Ada 002
text-embedding-ada-002 OpenAI's legacy text embedding model (1536 dimensions). Billed by input tokens via POST /v1/embeddings.
OpenAI
ImageGPT Image 2.5 Flare
gpt-image-2.5-flare OpenAI's fastest model for high-quality, everyday image generation. It accepts text and image inputs and produces image outputs. Quality supports low, medium, high, xhigh, max, and auto. Use the Images API or the Responses API image generation tool.
OpenAI
ImageGPT Image 2.5 Sunburst
gpt-image-2.5-sunburst OpenAI's GPT Image 2.5 Sunburst generates and edits images from text and image inputs. Use it for workflows where editing precision matters most. Quality supports low, medium, high, xhigh, max, and auto. Use the Images API or the Responses API image generation tool.
OpenAI
ImageGPT Image 2
gpt-image-2 OpenAI's high-fidelity image generation and editing model for the dedicated Images API.
OpenAI
LLMGPT-6 Astra
gpt-6-astra OpenAI's most capable model for demanding end-to-end work: reasoning, coding, computer use, research, and document creation. reasoning.effort supports low, medium, high, xhigh, and max (none is not supported). Tool calling should use the Responses API; Chat Completions remains available for text-only requests.
OpenAI
LLMGPT-6 Sol
gpt-6-sol A GPT-6 model for complex coding and agentic workflows. reasoning.effort supports none, low, medium (default), high, xhigh, and max. Built-in tools and function calling should use the Responses API; Chat Completions supports function calling only when reasoning_effort is none.
OpenAI
LLMGPT-6.1 Sol
gpt-6.1-sol An upgrade to GPT-6 Sol with near-Astra performance for complex coding, computer use, and professional work. reasoning.effort supports low, medium (default), high, xhigh, and max (none and minimal are not supported). Tool calling requires the Responses API; Chat Completions supports requests without tools.
OpenAI
LLMGPT-6 Luna
gpt-6-luna OpenAI's most efficient GPT-6 model for focused, high-volume tasks. reasoning.effort supports none, low, medium (default), high, xhigh, and max. Built-in tools and function calling should use the Responses API; Chat Completions supports function calling only when reasoning_effort is none.
OpenAI
LLMGPT-5.2
gpt-5.2 A frontier GPT model with adaptive reasoning for coding, agents, and long-context work. Adaptive reasoning allocates computation according to task difficulty for stronger agent and long-context performance.
OpenAI
LLMGPT-5.4
gpt-5.4 A flagship GPT model for advanced reasoning, coding, multimodal input, and professional workflows. It unifies general GPT and coding capabilities for large-context professional work and tool-driven execution.
OpenAI
LLMGPT-5.4 mini
gpt-5.4-mini A faster GPT-5.4 variant for high-throughput reasoning, coding, and multimodal applications. The smaller variant keeps strong reasoning and coding capabilities while improving speed and throughput.
OpenAI
LLMGPT-5.4 nano
gpt-5.4-nano A compact GPT model optimized for low-latency, high-volume, and cost-sensitive tasks. It is tuned for speed-critical classification, extraction, routing, and other high-volume lightweight tasks.
OpenAI
LLMGPT-5.5
gpt-5.5 A frontier GPT model for complex professional work with stronger reasoning and reliability. It strengthens reliability and token efficiency on difficult professional tasks while retaining long-context support.
OpenAI
LLMGPT-5.6
gpt-5.6 A GPT-5.6 family model for advanced reasoning, coding, and general agent workflows.
OpenAI
LLMGPT-5.6 Sol
gpt-5.6-sol The GPT-5.6 flagship for demanding reasoning, coding, and multi-step agent tasks. The flagship tier is particularly strong at command-line work and complex multi-step coding workflows.
OpenAI
LLMGPT-5.6 Terra
gpt-5.6-terra A balanced GPT-5.6 model for everyday coding, reasoning, and agentic applications. The balanced tier sits between flagship quality and cost efficiency for everyday coding, reasoning, and agents.
OpenAI
LLMGPT-5.6 Luna
gpt-5.6-luna A fast, efficient GPT-5.6 model for chat, classification, and lightweight agents. The efficient tier targets high-volume chat, classification, and lightweight agent workflows with low latency.
StepFun
LLMStep 3.7 Flash
step-3.7-flash An efficient multimodal StepFun model for native image and video understanding. Its multimodal mixture-of-experts design combines a large language backbone with native image and video understanding.
Tencent
LLMHy4 preview
hy4-preview Tencent's next-generation productivity model with significantly enhanced Agent and complex task execution capabilities. Hy4 preview features 770B total parameters with 49B active parameters and is optimized for agentic, coding, and productivity scenarios, with stronger capabilities in understanding, planning, tool use, and sustained execution for complex tasks.
Tencent
LLMHy3
hy3 A Tencent mixture-of-experts model for configurable reasoning and production agents. Configurable reasoning effort lets production systems trade latency for depth across agentic workloads.
Tencent
LLMHy3 preview
hy3-preview A high-efficiency Tencent preview model for agent workflows and production evaluation. The preview offers configurable reasoning levels for efficient evaluation and deployment of agent workflows.
xAI (Grok)
ImageGrok Imagine Image 2.0
grok-imagine-image-2.0 xAI's Imagine Image 2.0 model for precise generation, editing, and multi-reference creative work. It follows detailed instructions, preserves identity across edits, and supports 1K/2K output.
xAI (Grok)
ImageGrok Imagine Image Quality
grok-imagine-image-quality A high-fidelity xAI model for fast image generation, editing, and reference-guided creation. Retires on 2026-11-02; requests are then served by Grok Imagine Image 2.0 at low quality.
xAI (Grok)
LLMGrok Build 0.1
grok-build-0.1 An xAI coding-specialty model for agentic software development. Supports reasoning at configurable effort levels. Optimized for sustained coding trajectories and tool-driven workflows.
xAI (Grok)
LLMGrok 4.20 Reasoning
grok-4.20-0309-reasoning A 1M-context xAI reasoning model for complex coding, agents, and knowledge work. Supports reasoning at configurable effort levels. Batch API available at 20% off standard pricing.
xAI (Grok)
LLMGrok 4.20 Non-Reasoning
grok-4.20-0309-non-reasoning A 1M-context xAI model for coding, agents, and knowledge work without extended reasoning. Same pricing as the reasoning variant. Batch API available at 20% off standard pricing.
xAI (Grok)
LLMGrok 4.7
grok-4.7 SpaceXAI's frontier model for coding, agentic tasks, and knowledge work. Reasoning effort supports low, medium, high (default), and xhigh. The Responses API always returns encrypted reasoning content, which later turns must send back unchanged.
xAI (Grok)
LLMGrok 4.6
grok-4.6 A frontier Grok model for long-running agents, coding, and knowledge work. It focuses on multi-step agents, ambitious interactive and visual work, and sustained coding trajectories.
xAI (Grok)
LLMGrok 4.5
grok-4.5 A general-purpose Grok model for coding, knowledge work, STEM, and demanding reasoning. It supports both reasoning and non-reasoning modes for agentic software and workflow tasks.
xAI (Grok)
LLMGrok 4.3
grok-4.3 A multimodal Grok reasoning model for agents, instruction following, and factual tasks. It combines visual input with reasoning for factual instruction following and agent-oriented applications.
Xiaomi Mimo
LLMMiMo V2 Omni
mimo-v2-omni A Xiaomi omni-modal model for unified text, image, audio, and video understanding.
Xiaomi Mimo
LLMMiMo V2 Pro
mimo-v2-pro A capable Xiaomi MiMo model for complex reasoning, coding, and agentic workloads.
Xiaomi Mimo
LLMMimo V2.5 Pro
mimo-v2.5-pro Xiaomi's flagship MiMo model for software engineering and long-horizon agent tasks. The flagship emphasizes general agent capability, complex software engineering, and long-horizon execution.
Xiaomi Mimo
LLMMiMo V2.5
mimo-v2.5 A cost-efficient omni-modal MiMo model for agentic and visual understanding workloads. The native omni-modal design aims for Pro-level agent performance with more economical inference and stronger visual understanding.
Xiaomi Mimo
LLMMiMo V2.6 Pro
mimo-v2.6-pro Xiaomi's flagship omni-modal MiMo model for professional, long-horizon, and complex workflows, including scientific research and cybersecurity.
Xiaomi Mimo
LLMMiMo V2.6 Flash
mimo-v2.6-flash A cost-efficient omni-modal MiMo model for high-frequency professional work and large-scale tasks.
Xiaomi Mimo
LLMMiMo V2.6 Pro UltraSpeed
mimo-v2.6-pro-ultraspeed The high-speed serving mode of MiMo V2.6 Pro. It keeps flagship quality and offers up to 20× output speed for latency-sensitive workloads.
Zhipu AI
ImageGLM Image
glm-image A Zhipu image model for text-guided generation and general visual creation.
Zhipu AI
LLMGLM-5
glm-5 A flagship open GLM model for systems design, coding, and long-horizon agents. The open foundation model targets complex systems design and production-grade programming across long-running agent workflows.
Zhipu AI
LLMGLM-5 Turbo
glm-5-turbo A fast GLM model optimized for real-world agent environments and efficient inference. It is deeply optimized for fast inference in real-world, tool-using agent environments.
Zhipu AI
LLMGLM-5.1
glm-5.1 A GLM model with stronger coding and sustained performance on long-horizon tasks. Its coding improvements focus on sustained, independent execution over tasks that extend far beyond short interactions.
Zhipu AI
LLMGLM-5.2
glm-5.2 A long-context GLM reasoning model for project-level software engineering and agents. The large-scale reasoning model is intended for project-level software engineering and long-horizon agent workflows.
Zhipu AI
LLMGLM-5.3
glm-5.3 A flagship GLM model for long-horizon coding, agents, and security review. It keeps the GLM-5.2 base and raises coding and agent performance through post-training.
Zhipu AI
LLMGLM-5.3-Flash
glm-5.3-flash A native multimodal GLM Flash model for visual coding, agents, and professional workflows. It outperforms GLM-5.2 at a fraction of the cost, with native image, video, and file understanding.
Zhipu AI
LLMGLM-5.3-FlashX
glm-5.3-flashx A high-throughput GLM-5.3-Flash serving SKU at up to 200 tokens/s. It keeps native image, video, and file understanding with the same 1M context window.
No matches found
Try another keyword or clear the current filters.
Catalog prices come from Gateway presets. Confirm current pricing and availability with the vendor.
Data and contributions
Found incorrect Model data?
Every model card links to its preset file. Fix the ID, context, modalities, or catalog pricing and open a PR.