[Bug] LLM generates duplicate consecutive response blocks

Resolved 💬 4 comments Opened Apr 8, 2026 by openclosure Closed Apr 12, 2026

Has been happening repeatedly. Model writes ~90% of a response, somehow gets cut off when almost finished, then regenerates an entirely new response to the same user message. After both responses, both are in-context.

Bug Description
Model double-responded

Environment Info

  • Platform: darwin
  • Terminal: ghostty
  • Version: 2.1.97
  • Feedback ID: 5daeebea-4fc2-4d05-9e00-bd0a6001b2fb

Errors

[{"error":"Error: Request was aborted.\n    at makeRequest (/$bunfs/root/src/entrypoints/cli.js:50:3447)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-04-08T23:01:49.212Z"},{"error":"Error: Request was aborted.\n    at PV_ (/$bunfs/root/src/entrypoints/cli.js:1245:11933)\n    at next (native:1:11)\n    at Kb7 (/$bunfs/root/src/entrypoints/cli.js:7767:9194)\n    at next (native:1:11)\n    at Tb7 (/$bunfs/root/src/entrypoints/cli.js:7772:11371)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-04-08T23:05:58.739Z"},{"error":"MaxFileReadTokenExceededError: File content (12835 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at HT7 (/$bunfs/root/src/entrypoints/cli.js:4309:12666)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-04-08T23:06:33.084Z"}]

View original on GitHub ↗

This issue has 4 comments on GitHub. Read the full discussion on GitHub ↗