mitsuhiko's bash-only Benchmark Yields Mixed Results

mitsuhiko reports mixed results from bash-only benchmarking: tool call and edit failures increase, and new Sonnet's image reading and transcript readability remain pain points.

2026-09-22 ~ 2026-09-22 · 2 related posts

• mitsuhiko's Bash-Only Benchmark Results Are Mixed, With More Tool-Call Failures — mitsuhiko · 2026-09-22
• Mitsuhiko: Claude Code struggles with transcripts and needs a read tool for images — mitsuhiko · 2026-09-22