Grok 4.7 scores just 24% on CritPt v2 physics benchmark, trailing Chinese models
teortaxesTex · x · 2026-10-28
Citing the Artificial Analysis CritPt v2 (Hard Subset) physics reasoning leaderboard, teortaxesTex notes that xAI's Grok 4.7 (xhigh) scored only 24% — "just non-competitive" even against broadly weaker Chinese base models, with DeepSeek V4, Kimi, GLM and Qwen variants all scoring above it.
The leaderboard is topped by GPT-6-Astra (max) at 71%, GPT-6.1 Sol (max) at 69%, GPT-6.1 Terra (max) at 65%, Claude Opus 5.5 at 61% and Gemini 4 at 53%. The author also questions whether Grok's weak showing is compute-dependence or a possible bug.
More from Models
- Abliterated models terrify me, but so do safety filters: paraschopra on AI's fairness asymmetry — paraschopra · 2026-10-28
- New Paper Shows Gemini-Found Softmax Bug Fix Still Fails Property Tests — brandondamos · 2026-10-28
- Reddit users complain ChatGPT's coding quality keeps getting worse — LSP-86 · 2026-10-28
- Cryptic insider tease sparks guesses: GPT-6.2 Sol or GPT-6.1 Astra? — CtrlAltDwayne · 2026-10-28
- ECI Metric Adopted by Anthropic, Google and Artificial Analysis — Jsevillamol · 2026-10-28
- 'The OpenAI Canyon' meme chart shows OpenAI models trailing on output tokens — steipete · 2026-10-28