METR chart shows AI benchmark progress nearly doubled, Anthropic Model 4 hits 178.6 AECI
scaling01 · x · 2026-11-06
scaling01 highlights METR's latest AECI trend chart, claiming his earlier prediction was vindicated: Anthropic's Model 4 scores 178.6 on AECI, an aggregate capability metric mixing internal and external benchmarks.
- METR's chart shows that as of October 2026, benchmark scores are growing at 1.9× their pre-October 2025 rate (95% CI 1.5×–2.8×) when internal models and benchmarks are included.
- The trend line accelerates from +15/year to +29/year, with internal early models like Opus 5.5, Mythos 5.1, and Model 2/3/4 appearing on the curve.
- The author concludes AI progress has roughly doubled, expects RSI and superintelligent models by 2027–2028, and argues bio progress is bottlenecked by automated labs rather than model capability.
More from Models
- Gemini fails a simple prompt: asked for one-loop knot hourglass, drew three — Accreddita · 2026-11-06
- Ampersand matches frontier coding models at 70% lower cost via co-coding training — niloofar_mire · 2026-11-06
- theshawwn: government employees stay quiet, so Claude's unintuitive behaviors go unreported — theshawwn · 2026-11-06
- 21-model eval: 57.2% overall pass rate, no model clearly leads at 82.2% — davidyin44 · 2026-11-06
- Nemotron 3.5 Super + Surge RL matches GPT-5.6 Terra across 12 benchmarks — echen · 2026-11-06
- DeepSeek V4.1 Flash hits 600+ tok/s on Isoquant, 8.7x faster than Claude Opus 5.5 — ycombinator · 2026-11-06