Users Across Platforms Report Sharp Astra Degradation, Suspect Silent Nerf
Astra earned high praise shortly after launch, but starting around Saturday, September 13, multiple users on Reddit and X reported a sharp performance drop, suspecting a silent nerf or compute reallocation; one user also reported similar degradation in Fable. As of the material's coverage window, there is no official response or changelog.
Confirmed
- Reddit user Optimal-Sea8310 (09-14) said Astra was smart at launch but began making basic errors even Sol wouldn't, becoming nearly unusable, with redo time exceeding actual work time; they speculated compute was being diverted.
- Economist Joachim Voth (09-15) complained Astra had solved two major problems and supervised full drafts and simulations the previous week—work he had publicly praised—but since Saturday it barely moves, like being "out of tokens."
- An indie developer (via amasad's repost, 09-15) said Astra on Replit is "completely unusable," with game-building quality clearly down, suspecting OpenAI nerfed the model.
- Designer Tom John (tomjohndesign, posting twice on 09-15/09-16, once naming OpenAI and once Google) said Astra is "no longer the same model" as last week and believes it was nerfed.
- Developer GainsdalfGwei (09-15) slammed the quantized/nerfed GPT-6 Astra in Codex: constant over-correction, gratuitous rebuilding of existing implementations, no common sense—"a monkey's paw."
- User RileyRalmuto (09-15) said both Fable and Astra have "gotten dumber," with more regressions and errors than since either launched.
- soumitrashukla9 (09-15) replied to Voth saying they hit similar glitches and suggested workarounds.
Unconfirmed
- Whether the model was actually quantized, nerfed, or had compute diverted remains user speculation; it could equally be a silent update or transient outage, with no official word.
- Posts disagree on which "Astra" is meant: some point to Google Project Astra/Gemini, others to OpenAI, and GainsdalfGwei explicitly names GPT-6 Astra in Codex; whether this is one product's degradation or several same-named products being conflated is unclear.
- No further evidence links Fable's degradation to Astra's.
Why it matters
- If confirmed, this fuels controversy over providers silently altering model capabilities post-launch and eroding user trust; users from diverse backgrounds (economist, designer, game developer) reporting similar degradation in the same window makes official response worth watching.
2026-09-14 ~ 2026-09-16 · 8 related posts
- Episode 1: GPT-6 Astra Hit by Silent Downgrade Claims; OpenAI Admits and Fixes(2026-09-11, 10 posts)
- Episode 2: Users Across Platforms Report Sharp Astra Degradation, Suspect Silent Nerf(2026-09-14, 8 posts)
Primary sources
- [source] Astra users report significant performance degradation days after launch — Optimal-Sea8310 · 2026-09-14
- Indie dev says OpenAI's Astra is 'absolutely useless' now, suspects silent nerf — amasad · 2026-09-15
- Users Report Fable and Astra Regressions: Worst Error Rate Since Launch — RileyRalmuto · 2026-09-15
- [source] Dev slams quantized Astra in Codex: overcorrects, lacks common sense, 'monkey's paw' — GainsdalfGwei · 2026-09-15
- [source] Astra went from workhorse to "out of tokens" overnight, researcher reports — joachim_voth · 2026-09-15
- Users Report Gemini Astra Suddenly Degrading: "Like a Slow-Clap Version of Itself" — soumitrashukla9 · 2026-09-15
- User claims OpenAI nerfed Astra: 'not the same model it was last week' — tomjohndesign · 2026-09-15
- Designer claims Google 'nerfed' Astra: 'this is not the same model it was last week' — tomjohndesign · 2026-09-16