Skip to content

Available Models

The platform aggregates mainstream models from multiple providers. The tables below are examples of commonly available models. For the exact model IDs, context lengths, and unit prices, refer to the console Models / Pricing page — they are adjusted along with upstream channels.

TIP

Use the GET /v1/models endpoint to fetch the list of models actually available to your token — this is the authoritative way to check whether a model is available.

Chat & text completion models

OpenAI series

Model IDContextVisionFunction callingNotes
gpt-4o128KFlagship multimodal, balances speed and quality
gpt-4o-mini128KLightweight and fast, for high-frequency simple tasks
o3200KDeep-reasoning flagship
o4-mini200KReasoning model, cost-effective
gpt-4.11MUltra-long context, strong coding

Anthropic Claude series

Model IDContextVisionFunction callingNotes
claude-opus-4200KTop-tier reasoning and writing
claude-sonnet-4200KBalanced capability and speed, recommended default
claude-haiku-3-5200KVery fast and light, for real-time interaction

Google Gemini series

Model IDContextVisionFunction callingNotes
gemini-2.5-pro1MUltra-long context, multimodal flagship
gemini-2.5-flash1MGreat value, very fast

DeepSeek series

Model IDContextVisionFunction callingNotes
deepseek-v364KStrong all-round ability, optimized for Chinese
deepseek-r164KDeep reasoning with chain-of-thought output

Qwen series

Model IDContextVisionFunction callingNotes
qwen-max32KAlibaba Cloud flagship, best-in-class Chinese
qwen-vl-max32KSpecialized vision understanding
qwen-long1MLong-document processing

Embedding models

Model IDDimensionsNotes
text-embedding-3-small1536General semantic search, cost-effective
text-embedding-3-large3072High-precision semantic understanding
text-embedding-ada-0021536Compatibility with older projects

Image generation models

Model IDMax resolutionNotes
dall-e-31792×1024By OpenAI, stable quality
gpt-image-11536×1024New-generation image model
flux-1-pro2048×2048Open-source ecosystem, rich detail
stable-diffusion-31024×1024Community ecosystem

Audio models

Model IDTypeNotes
tts-1TTSStandard quality, low latency
tts-1-hdTTSHD quality
whisper-1STTMultilingual recognition, 99 languages

Video generation models

Model IDMax durationNotes
sora-220sBy OpenAI, high quality
kling-v210sOptimized prompts, natural motion
wan-2.18sOpen-source video model

Note on model updates

The platform continually onboards the latest models; the tables above are for reference only. The real-time data on the console Pricing / Models page is authoritative.

OpenAI-compatible · Multimodal AI gateway