Artificial Analysis Intelligence Index v5: nine benchmarks across four equally weighted pillars
ArtificialAnlys · x · 2026-10-28
Artificial Analysis published the v5 weights for its Intelligence Index, splitting 25% each across Knowledge Work (AA-Briefcase 15%, AutomationBench-AA 5%, GDP.pdf 5%), Coding (FrontierCode 2.0 15%, Terminal-Bench 4.0 10%), Science (Terminal-Bench-Science 15%, CritPt v2 hard subset 10%), and General (AA-Omniscience v1.1 15%, HLE-Diamond 10%).
Key findings from the thread:
- Median output tokens per task jumped from 35k (v4.3) to 59k; agentic evals hold 55% of weight but consume 88% of tokens.
- Effort settings matter more: Claude Opus 5.5 scores 35 at low effort vs 58 at max ($0.98 vs $10.04 per task); GPT-6.1 Sol's spread doubled to 20 points.
- Cost gap is huge: Opus 5.5 max costs $10.04/task, GPT-6.1 Sol max $1.27 while scoring within 4 points.
- Open weights fell further behind: leader Kimi K3.1 Pro (weights unreleased) scores 46, 12 points off the top; released-weights DeepSeek V4.1 Flash 1017 and GLM-5.3 top out at 32, below Claude Haiku 5.5.
- No model sweeps all nine evals; Opus 5.5 leads three, with five other models leading the rest.
More from Models
- Debate: Claude isn't 'lobotomized' — inference stack updates may shift old-model outputs — JasonBotterill · 2026-10-28
- AI labs near IPO, yet LLMs still produce 'astonishingly terrible prose' by default — GabGarrett · 2026-10-28
- Rebuttal: Kimi 3.1 Nears Fable on Frontier Code 2, Contradicting 'China Models Crushed' Claim — bookwormengr · 2026-10-28
- Open model Flint hits #2 trending on Hugging Face, team outlines next steps — josh_tobin_ · 2026-10-28
- Artificial Analysis Intelligence Index v5 widens open-weight gap from 3 to 12 points — ResearchCrafty1804 · 2026-10-28
- YacineMTB rails: even the best LLMs are 'complete dogshit' and he's tired of it — yacineMTB · 2026-10-28