Skip to content

v0.1.2 - Fix Claude Token Limit

Choose a tag to compare

@mmcky mmcky released this 01 Oct 06:22
· 109 commits to main since this release

Critical Bug Fix 🐛

Fixed max_tokens API error for Claude 3.5 Sonnet

This release resolves the critical error encountered when reviewing lectures:

Error code: 400 - {'type': 'error', 'error': {'type': 'invalid_request_error', 
'message': 'max_tokens: 16000 > 8192, which is the maximum allowed number of 
output tokens for claude-3-5-sonnet-20241022'}}

What Changed

  • ✅ Reduced max_tokens from 16000 to 8192 for Claude 3.5 Sonnet compatibility
  • ✅ Added comprehensive LLM model documentation with token limits
  • ✅ Added troubleshooting guide for token limit errors
  • ✅ Enhanced trigger to support @mentions in issue bodies (not just comments)

New Documentation

  • Model Options Guide: Compare Claude, GPT-4, and Gemini models
  • Token Limits: Reference table for all supported models
  • Troubleshooting: Solutions for token limit errors

Model Recommendations

  • Claude 3.5 Sonnet (default): 8,192 tokens - Best for most lectures
  • GPT-4o: 16,384 tokens - For very long lectures
  • Gemini 1.5 Pro: 8,192 tokens - Cost-effective alternative

Upgrade Notes

Users on v0.1.0 or v0.1.1 should upgrade immediately if encountering token limit errors.

The floating tag v0.1 now points to this release for automatic patch updates.


Full Changelog: v0.1.1...v0.1.2