Artificial Analysis ships CritPt v2 grading server, regrades all models with max 1-point shifts
ArtificialAnlys · x · 2026-11-04
Artificial Analysis partnered with the CritPt authors to finalize the CritPt v2 grading server, available to the community via its API, and folded the robustness improvements into Intelligence Index v5.0.1.
- All models were regraded: small score increases for most
- Frontier model rankings unchanged, scores moving by at most 1 point
- The chart shows GPT-6 leading at 83.1%, followed by GPT-5.1 (High) at 82.6% and Gemini 4 Canvas (High) at 82.3%; Z 2B King (High) trails at 10.6%
More from Models
- Legora CTO on frontier models that can understand and modify their own weights — garrytan · 2026-11-04
- New research lets models compose their own computational graphs — micahgoldblum · 2026-11-04
- LoL at matched FLOPs rivals dense transformers at 1B params, router probe hits 0.95 AUROC — micahgoldblum · 2026-11-04
- LoL architecture reorders its own forward pass at runtime, mimicking algorithmic composition — micahgoldblum · 2026-11-04
- Grok Bot weekly quota 30% used within 9 hours of reset, user asks for bigger pool — nima_owji · 2026-11-04
- Ethan Mollick's test: just 5 comments through Pangram flag 100% of X AI-slop accounts — emollick · 2026-11-04