AI benchmark news
AI benchmark news: new results and evaluations of AI models, attributed to whoever ran them.

Artificial Analysis Intelligence Index v5, October 27, 2026: Claude Opus 5.5 leads at 58
Artificial Analysis rebuilt its index on Oct 27, 2026: Claude Opus 5.5 leads at 58, GPT-6 Astra 57, GPT-6.1 Sol 55.

Vera Rubin NVL72 benchmark: AA-AgentPerf results on October 15, 2026
Artificial Analysis, October 15, 2026: Vera Rubin NVL72 serves up to ~5.7x GB300 NVL72's throughput per MW.

Epoch AI InnovationEval results on October 7, 2026: agents reach 15%
Epoch AI's October 7, 2026 InnovationEval: the best in-scope agent result reached 15% of a human method's gains.

Gemini 4 Argon benchmarks: Artificial Analysis results, Sep 30, 2026
Gemini 4 Argon scored 53 on Artificial Analysis on Sept 30, 2026, and 50 on its rebuilt Intelligence Index v5.

Claude Opus 5.5 tops the Intelligence Index on September 22, 2026
On September 22, 2026, Artificial Analysis scored Claude Opus 5.5 at 58 on its Intelligence Index at max effort.

Grok 4.7 benchmarks: xAI's reported results, September 21, 2026
xAI released Grok 4.7 on September 21, 2026 and reported its benchmarks, including 71.0% on DeepSWE v1.1.

ARC Prize publishes GPT-6 Astra ARC-AGI-3 results on September 3, 2026
On September 3, 2026, ARC Prize reported GPT-6 Astra at 62.7% on ARC-AGI-3 Semi-Private with its Standard harness.
Scores come from the lab’s model card or technical report, or from the benchmark maintainer’s own leaderboard, always with the benchmark version and the date. Published scores are collected on the LLM leaderboard, and what SWE-bench Verified, GPQA Diamond, ARC-AGI or Humanity’s Last Exam measure is covered in our benchmark explainers.
