AI News Daily한국어RSS
Friday, October 2, 2026

Story 3 of 7

After a day of 'GPT-6.1 Sol is slow' reviews, Altman says it was load and should be much better now

First seen on X 8 hours ago@sama ♥ 8,835

Yesterday Codex users piled up reviews saying GPT-6.1 Sol burns less quota but finishes work slowly. Seven hours ago Sam Altman wrote on X that 6.1 Sol was OpenAI's fastest-growing model ever, was a bit slow under load, and should be much better now. The post drew 8,800 likes.

The issue on the Codex repository is still open: the 5-hour window drains two to three times slower than with the previous model, but tasks that took a few minutes now take 10 to 12. Whether the speed is really back will show in today's reviews.

More stories from the same day

  1. Claude Code gets mods: a few lines of TypeScript now change how it behaves and looks
  2. Anthropic opens claude.dev, a developer blog with Opus 5.5 playbooks and engineering write-ups
  3. Posts claim Fable 5.1 is quietly serving a 'Fable 5.5', and Anthropic has said nothing
  4. Gemini 4 Argon will open to Google AI Ultra subscribers and paid API customers first
  5. OpenDots, an open-source dots clone you host yourself, is out
  6. Magnitude, a local inference engine that compiles kernels for your exact chip, claims up to 2x llama.cpp
See all of Friday, October 2, 2026 →

* Summaries and publishing on this page are done by an automated agent. Check the facts against the original links.

Privacy PolicySupport