Bug Hunt Bench: GPT-6.2 Astra max tops at 62.7, beats Fable 5.2 at half the cost
PawelHuryn · x · 2026-11-19
Pawel Huryn shared all effort levels for GPT-6.2 Astra and GPT-6.2 Sol on Bug Hunt Bench (planted bugs fixed, blind-verified), plotted score vs. cost on a log scale:
- GPT-6.2 Astra·max leads at 62.7, with xhigh 60.2, high 59.6, medium 50.3 — all in the "cheap and strong" zone around $20-40
- Fable 5.2·max scores 58 but costs roughly $150-200, the most expensive run
- GPT-6.2 Sol max hits 54.7 (xhigh 49); last-gen GPT-6.1 Sol max only 44.3, showing a clear generational jump
- Opus 5.5 max 41.7, xhigh 36, high 31.7 at $10-60 — its score-cost curve is dominated by the GPT-6.2 family
- GPT-6 Astra max 45 sits between generations
Takeaway: GPT-6.2 Astra/Sol beat Fable 5.2 and Opus 5.5 across nearly every price tier, with the widest gaps at high effort levels.
More from Models
- Gemini 4's element naming mocked: Carbon builds on Argon, Neon on Hydrogen — teortaxesTex · 2026-11-19
- Rumor: Gemini 4 is a 12T-param MoE with 500B active; ChatGPT translation fumbles it — teortaxesTex · 2026-11-19
- Pangram false positives are nearly nonexistent — Bender's high-FP prediction fails a cash-bounty test — FlorianGallwitz · 2026-11-19
- Abacus.AI open-sources Smaug-Uncensored, cutting Qwen3.8-27B refusal rate from 45% to 2% — bindureddy · 2026-11-19
- Dev observes Google's model launch pace has dramatically sped up since Argon — cgarciae88 · 2026-11-19
- Amazon's in-house web search agent takes #1 on Search Arena with score 1266 — arena · 2026-11-19