token

[HSFZXN]

Git provider (optional)

None

System Info (optional)

No response

Issues details

配置中模型设置的token上限是多少

[qodo-free-for-open-source-projects]

💡 Auto-generated documentation-based answer:

Based on the documentation, there are several token-related configurations:

1. Default Model Token Limits: The default model (GPT-5) has a context window that can handle most PRs up to 600 lines of code in a single call.

2. Custom Model Token Configuration: You can set a custom token limit using:

   [config]
   custom_model_max_tokens = <your_value>

   For example, when using Ollama models, you might set it to 128000.

3. Model-Specific Limits:

   • For Ollama: Default context window is 2048 tokens, but can be overridden with OLLAMA_CONTEXT_LENGTH environment variable
   • For different models, the limits vary (e.g., GPT-4 has 8000 tokens, GPT-32K has larger context)

4. PR Compression Settings:

   [config]
   max_model_tokens = 32000  # Default for chunking large PRs

To get a more specific answer, please clarify:

• Which model are you using (OpenAI, Gemini, Claude, Ollama, etc.)?
• Are you asking about input token limits or output token limits?
• Are you using the open-source PR-Agent or Qodo Merge?

Relevant Sources:

• https://docs.qodo.ai/qodo-documentation/qodo-merge/usage-guide/changing_a_model#ollama
• https://docs.qodo.ai/qodo-documentation/qodo-merge/usage-guide/changing_a_model#custom-models
• https://docs.qodo.ai/qodo-documentation/qodo-merge/tools/improve#how-many-code-suggestions-are-generated
• https://docs.qodo.ai/qodo-documentation/qodo-merge/usage-guide/additional_configurations#performance-optimization