Models

Access 38+ leading Chinese AI models through a single API. All models are accessible via the same OpenAI-compatible endpoint.

💡 Model selection guide
Use GET /v1/models to list all available models programmatically. For general tasks, we recommend starting with DeepSeek V4 Pro or Qwen 3.5 397B.

DeepSeek Models

DeepSeek V4 Pro Flagship

DeepSeek's most capable model. Exceptional at complex reasoning, math, and coding tasks.

💰 $1.65/$3.30 per 1M 📄 64K ctx 🎯 Reasoning
DeepSeek V4 Flash Fast

Optimized for speed. Great for high-throughput applications with good quality.

💰 $0.27/$0.41 per 1M 📄 64K ctx ⚡ Fast
DeepSeek V3.2 Stable

Previous generation flagship with proven reliability and broad capability.

💰 $0.55/$0.83 per 1M 📄 64K ctx 💪 General
DeepSeek V3 Economical

Cost-effective general-purpose model. Ideal for most standard applications.

💰 $0.27/$0.41 per 1M 📄 64K ctx 💰 Budget
DeepSeek R1 Reasoning

Dedicated reasoning model with chain-of-thought. Excellent for math and logic.

💰 $0.55/$2.19 per 1M 📄 64K ctx 🧠 CoT

Qwen (Alibaba) Models

Qwen 3 Max Preview Next-Gen

Alibaba's next-generation flagship with breakthrough performance.

💰 $0.41/$1.24 per 1M 📄 128K ctx 🏆 Top-tier
Qwen 3.5 397B-A17B MoE Giant

Massive MoE architecture: 397B total params, 17B active. Outstanding performance.

💰 $0.55/$1.10 per 1M 📄 128K ctx 🚀 MoE
Qwen Max Proven

Battle-tested flagship model. Reliable performance across all tasks.

💰 $0.27/$0.82 per 1M 📄 32K ctx ✅ Stable
Qwen Plus Balanced

Best price-performance ratio. Recommended for most production workloads.

💰 $0.14/$0.41 per 1M 📄 128K ctx 💰 Value
Qwen Turbo Fast

Ultra-fast responses for latency-sensitive applications.

💰 $0.07/$0.14 per 1M 📄 128K ctx ⚡ Lowest cost
Qwen 3 Coder Plus Code

Specialized for code generation, review, and debugging across 50+ languages.

💰 $0.27/$0.82 per 1M 📄 128K ctx 💻 Code
Qwen Long Long Context

Designed for very long documents. Process entire codebases or books.

💰 $0.07/$0.27 per 1M 📄 10M ctx 📖 Long ctx
Qwen VL Max Vision

Image understanding and visual question answering at the highest quality.

💰 $0.41/$1.24 per 1M 👁️ Vision 🎨 Multi-modal
Qwen VL Plus Vision Lite

Cost-effective image understanding. Great for OCR and image analysis.

💰 $0.14/$0.41 per 1M 👁️ Vision 💰 Budget
Qwen Audio Turbo Audio

Speech and audio understanding. Transcribe and analyze audio inputs.

💰 $0.14/$0.41 per 1M 🎧 Audio 🎤 Multi-modal

Other Top Models

Kimi K2.7 Code Code Expert

Moonshot AI's specialized coding model. Excellent at code generation and review.

💰 $0.68/$2.06 per 1M 📄 128K ctx 💻 Code
GLM-5.2 Zhipu

Zhipu AI's latest flagship. Strong Chinese language capabilities.

💰 $1.37/$4.40 per 1M 📄 128K ctx 🇨🇳 Chinese
MiniMax M2.5 Versatile

General-purpose model with excellent multilingual performance and creative writing.

💰 $0.55/$1.10 per 1M 📄 128K ctx 🌐 Multilingual
Step 3.5 Flash StepFun

StepFun's fast inference model. Good balance of speed and quality.

💰 $0.27/$0.55 per 1M 📄 32K ctx ⚡ Fast
LongCat 2.0 Long Input

Meituan's long-context model for processing large documents and codebases.

💰 $0.27/$0.55 per 1M 📄 1M ctx 📖 Long ctx
Nex N2 Pro NexAGI

Professional-grade model with strong analytical and summarization capabilities.

💰 $0.41/$1.10 per 1M 📄 64K ctx 📊 Analysis

Qwen 3.5 Variant Models

Multiple sizes available for different cost/performance needs:

Qwen 3.5 122B-A10BMoE

122B params, 10B active. Good mid-range option.

💰 $0.27/$0.55 per 1M📄 128K ctx
Qwen 3.5 35B-A3BEfficient

35B params, 3B active. Ultra-efficient for simple tasks.

💰 $0.07/$0.14 per 1M📄 128K ctx
Qwen 3.5 27BDense

27B dense model. Consistent performance for complex tasks.

💰 $0.14/$0.27 per 1M📄 128K ctx
Qwen 3.5 9BLight

9B dense model. Fast and economical for lightweight tasks.

💰 $0.04/$0.07 per 1M📄 128K ctx
Qwen 3.5 4BNano

4B model. Ultra-fast, lowest cost for simple classification and extraction.

💰 $0.02/$0.04 per 1M📄 128K ctx
Qwen 3.6 27BLatest

Latest Qwen 3.6 generation. Improved instruction following.

💰 $0.14/$0.27 per 1M📄 128K ctx

Embedding & Reranking Models

Qwen3 Embedding 8BEmbedding

High-quality embedding model for semantic search and RAG. 8B parameters.

💰 $0.07/$0.07 per 1M📈 4096 dims
Qwen3 Embedding 4BEmbedding

Lightweight embedding model. Good quality with lower latency.

💰 $0.04/$0.04 per 1M📈 2560 dims
BGE Reranker V2 M3Reranker

Cross-encoder reranking model for improving search result relevance.

💰 $0.04/$0.04 per 1M🔍 Reranking

🎁 Try Any Model Free

Use code DEVSTARTER20 to get ¥20 free credit — test any model before committing.

Start Free →