> ## Documentation Index > Fetch the complete documentation index at: https://aisa.one/docs/llms.txt > Use this file to discover all available pages before exploring further. # AIsa CN-LLM Route > Route Chinese-language prompts to Chinese LLM families through AIsa. [View on GitHub ->](https://github.com/AIsa-team/agent-skills/tree/main/cn-llm) **Chinese LLM routing through AIsa.** Send Chinese-language tasks to Qwen, DeepSeek, GLM, Baichuan, and related model families. ## Install First, install the AIsa CLI if you have not already: ```bash theme={null} npm install -g @aisa-one/cli ``` Then install the skill: ```bash theme={null} aisa skills install cn-llm ``` ## What can agents do with it? Route Chinese-language tasks to strong Chinese LLMs. Choose between Qwen, DeepSeek, GLM, Baichuan, and related models. Switch between Chinese and global model families. Give agents practical routing heuristics. ## 🔥 What You Can Do ### Intelligent Chat ``` "Use Qwen to answer Chinese questions, use DeepSeek for coding" ``` ### Deep Reasoning ``` "Use DeepSeek-R1 for complex reasoning tasks" ``` ### Code Generation ``` "Use DeepSeek-Coder to generate Python code with explanations" ``` ### Long Text Processing ``` "Use Qwen-Long for ultra-long document summarization" ``` ### Model Comparison ``` "Compare response quality between Qwen-Max and DeepSeek-V3" ``` ## Supported Models ### Qwen (Alibaba) | Model | Input Price | Output Price | Features | | ------------------------------ | ----------- | ------------ | --------------------------- | | qwen3-max | \$1.37/M | \$5.48/M | Most powerful general model | | qwen3-max-2026-01-23 | \$1.37/M | \$5.48/M | Latest version | | qwen3-coder-plus | \$2.86/M | \$28.60/M | Enhanced code generation | | qwen3-coder-flash | \$0.72/M | \$3.60/M | Fast code generation | | qwen3-coder-480b-a35b-instruct | \$2.15/M | \$8.60/M | 480B large model | | qwen3-vl-plus | \$0.43/M | \$4.30/M | Vision-language model | | qwen3-vl-flash | \$0.86/M | \$0.86/M | Fast vision model | | qwen3-omni-flash | \$4.00/M | \$16.00/M | Multimodal model | | qwen-vl-max | \$0.23/M | \$0.57/M | Vision-language | | qwen-plus-2025-12-01 | \$1.26/M | \$12.60/M | Plus version | | qwen-mt-flash | \$0.168/M | \$0.514/M | Fast machine translation | | qwen-mt-lite | \$0.13/M | \$0.39/M | Lite machine translation | ### DeepSeek | Model | Input Price | Output Price | Features | | ---------------- | ----------- | ------------ | ------------------------------- | | deepseek-r1 | \$2.00/M | \$8.00/M | Reasoning model, supports Tools | | deepseek-v3 | \$1.00/M | \$4.00/M | General chat, 671B parameters | | deepseek-v3-0324 | \$1.20/M | \$4.80/M | V3 stable version | | deepseek-v3.1 | \$4.00/M | \$12.00/M | Latest Terminus version | > **Note**: Prices are in M (million tokens). Model availability may change, see [console.aisa.one/pricing](https://console.aisa.one/pricing) for the latest list. ## Quick Start ```bash theme={null} export AISA_API_KEY="your-key" ``` ## API Endpoints ### OpenAI Compatible Interface ``` POST https://api.aisa.one/v1/chat/completions ``` #### Qwen Example ```bash theme={null} curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "qwen3-max", "messages": [ {"role": "system", "content": "You are a professional Chinese assistant."}, {"role": "user", "content": "Please explain what a large language model is?"} ], "temperature": 0.7, "max_tokens": 1000 }' ``` #### DeepSeek Example ```bash theme={null} # DeepSeek-V3 general chat (671B parameters) curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "deepseek-v3", "messages": [{"role": "user", "content": "Write a quicksort algorithm in Python"}], "temperature": 0.3 }' # DeepSeek-R1 deep reasoning (supports Tools) curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "deepseek-r1", "messages": [{"role": "user", "content": "A farmer needs to cross a river with a wolf, a sheep, and a cabbage. The boat can only carry the farmer and one item at a time. If the farmer is not present, the wolf will eat the sheep, and the sheep will eat the cabbage. How can the farmer safely cross?"}] }' # DeepSeek-V3.1 Terminus latest version curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "deepseek-v3.1", "messages": [{"role": "user", "content": "Implement an LRU cache with get and put operations"}] }' ``` #### Qwen3 Code Generation Example ```bash theme={null} curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "qwen3-coder-plus", "messages": [{"role": "user", "content": "Implement a thread-safe Map in Go"}] }' ``` #### Parameter Reference | Parameter | Type | Required | Description | | ------------- | ------- | -------- | -------------------------------- | | `model` | string | Yes | Model identifier | | `messages` | array | Yes | Message list | | `temperature` | number | No | Randomness (0-2, default 1) | | `max_tokens` | integer | No | Maximum tokens to generate | | `stream` | boolean | No | Stream output (default false) | | `top_p` | number | No | Nucleus sampling parameter (0-1) | #### Response Format ```json theme={null} { "id": "chatcmpl-xxx", "object": "chat.completion", "created": 1234567890, "model": "qwen-max", "choices": [ { "index": 0, "message": { "role": "assistant", "content": "A large language model (LLM) is a deep learning-based..." }, "finish_reason": "stop" } ], "usage": { "prompt_tokens": 30, "completion_tokens": 150, "total_tokens": 180, "cost": 0.001 } } ``` ### Streaming Output ```bash theme={null} curl -X POST "https://api.aisa.one/v1/chat/completions" \ -H "Authorization: Bearer $AISA_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "qwen-plus", "messages": [{"role": "user", "content": "Tell a Chinese folk story"}], "stream": true }' ``` Returns Server-Sent Events (SSE) format: ``` data: {"id":"chatcmpl-xxx","choices":[{"delta":{"content":"Once"}}]} data: {"id":"chatcmpl-xxx","choices":[{"delta":{"content":" upon"}}]} ... data: [DONE] ``` ## Python Client ### CLI Usage ```bash theme={null} # Qwen chat python3 scripts/cn_llm_client.py chat --model qwen3-max --message "Hello, please introduce yourself" # Qwen3 code generation python3 scripts/cn_llm_client.py chat --model qwen3-coder-plus --message "Write a binary search algorithm" # DeepSeek-R1 reasoning python3 scripts/cn_llm_client.py chat --model deepseek-r1 --message "Which is larger, 9.9 or 9.11? Please reason in detail" # DeepSeek-V3 chat python3 scripts/cn_llm_client.py chat --model deepseek-v3 --message "Tell a story" --stream # With system prompt python3 scripts/cn_llm_client.py chat --model qwen3-max --system "You are a classical poetry expert" --message "Write a poem about plum blossoms" # Model comparison python3 scripts/cn_llm_client.py compare --models "qwen3-max,deepseek-v3" --message "What is quantum computing?" # List supported models python3 scripts/cn_llm_client.py models ``` ### Python SDK Usage ```python theme={null} from cn_llm_client import CNLLMClient client = CNLLMClient() # Uses AISA_API_KEY environment variable # Qwen chat response = client.chat( model="qwen3-max", messages=[{"role": "user", "content": "Hello!"}] ) print(response["choices"][0]["message"]["content"]) # Qwen3 code generation response = client.chat( model="qwen3-coder-plus", messages=[ {"role": "system", "content": "You are a professional programmer."}, {"role": "user", "content": "Implement a singleton pattern in Python"} ], temperature=0.3 ) # Streaming output for chunk in client.chat_stream( model="deepseek-v3", messages=[{"role": "user", "content": "Tell a story about an idiom"}] ): print(chunk, end="", flush=True) # Model comparison results = client.compare_models( models=["qwen3-max", "deepseek-v3", "deepseek-r1"], message="Explain what machine learning is" ) for model, result in results.items(): print(f"{model}: {result['response'][:100]}...") ``` ## Use Cases ### 1. Chinese Content Generation ```python theme={null} # Copywriting response = client.chat( model="qwen3-max", messages=[ {"role": "system", "content": "You are a professional copywriter."}, {"role": "user", "content": "Write a product introduction for a smart watch"} ] ) ``` ### 2. Code Development ```python theme={null} # Code generation and explanation response = client.chat( model="qwen3-coder-plus", messages=[{"role": "user", "content": "Implement a thread-safe Map in Go"}] ) ``` ### 3. Complex Reasoning ```python theme={null} # Mathematical reasoning response = client.chat( model="deepseek-r1", messages=[{"role": "user", "content": "Prove: For any positive integer n, n³-n is divisible by 6"}] ) ``` ### 4. Visual Understanding ```python theme={null} # Image understanding response = client.chat( model="qwen3-vl-plus", messages=[ {"role": "user", "content": [ {"type": "text", "text": "Describe the content of this image"}, {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}} ]} ] ) ``` ### 5. Model Routing Strategy ```python theme={null} MODEL_MAP = { "chat": "qwen3-max", # General chat "code": "qwen3-coder-plus", # Code generation "reasoning": "deepseek-r1", # Complex reasoning "vision": "qwen3-vl-plus", # Visual understanding "fast": "qwen3-coder-flash", # Fast response "translate": "qwen-mt-flash" # Machine translation } def route_by_task(task_type: str, message: str) -> str: model = MODEL_MAP.get(task_type, "qwen3-max") return client.chat(model=model, messages=[{"role": "user", "content": message}]) ``` ## Error Handling Errors return JSON with `error` field: ```json theme={null} { "error": { "code": "model_not_found", "message": "Model 'xxx' is not available" } } ``` Common error codes: * `401` - Invalid or missing API Key * `402` - Insufficient balance * `404` - Model not found * `429` - Rate limit exceeded * `500` - Server error ## Pricing | Model | Input (\$/M) | Output (\$/M) | | ----------------- | ------------ | ------------- | | qwen3-max | \$1.37 | \$5.48 | | qwen3-coder-plus | \$2.86 | \$28.60 | | qwen3-coder-flash | \$0.72 | \$3.60 | | qwen3-vl-plus | \$0.43 | \$4.30 | | deepseek-v3 | \$1.00 | \$4.00 | | deepseek-r1 | \$2.00 | \$8.00 | | deepseek-v3.1 | \$4.00 | \$12.00 | > Price unit: \$ per Million tokens. Each response includes `usage.cost` and `usage.credits_remaining`. ## Get Started 1. Register at [aisa.one](https://aisa.one) 2. Get API Key 3. Top up (pay-as-you-go) 4. Set environment variable: `export AISA_API_KEY="your-key"` ## Full API Reference See [API Reference](https://aisa.one/docs/api-reference) for complete endpoint documentation. ## Get started 1. Sign up at [aisa.one](https://aisa.one) (new accounts start with \$2 free credit). 2. Generate an API key from the console. 3. Set your key and install the skill: ```bash theme={null} export AISA_API_KEY="your-key" npm install -g @aisa-one/cli aisa skills install cn-llm ``` 4. Start a new agent session so the runtime loads the updated skill instructions. ## Related Overview of Chinese model families in AIsa. General-purpose model routing across providers. Browse supported model IDs.