Artificial Analysis article Announcing Artificial Analysis Intelligence Index v5, dated October 27, 2026
Artificial Analysis's Intelligence Index v5 announcement, captured 2026-10-28. Image: AppFor.ai screenshot of Artificial Analysis's page. Image source ↗

What happened

On October 27, 2026, Artificial Analysis published version 5 of its Intelligence Index, a 0 to 100 composite of nine evaluations. It calls v5 its biggest upgrade since v4 launched in January, and says 60% of the index now rests on private test sets or held-out answers. It launched a separate Progress Index v1 for tracking capability over time.

  • Four categories at 25% each: Knowledge Work, Coding, Science and General.
  • New: FrontierCode 2.0 (15%) and Terminal-Bench-Science (15%).
  • Upgraded: HLE-Diamond, AA-Omniscience v1.1 and a 40-question hard subset of CritPt v2.
  • Removed: GDPval-AA v2.1, SciCode and AA-LCR v1.1.

Key results

All figures are Artificial Analysis's, at the effort level shown.

  • Claude Opus 5.5 (max) 58, GPT-6 Astra (max) 57, GPT-6.1 Sol (max) 55; Gemini 4 Argon (high) 50.
  • Top open weights model: Kimi K3.1 Pro at 46, weights not yet released, 12 points behind against a 3-point gap in v4.3. Among released weights, DeepSeek V4.1 Flash 1017 and GLM-5.3 lead at 32.
  • Cost per index task among the top 10: from $0.96 for GPT-6.1 Terra (max) to $13.04 for Claude Fable 5.1 (max); Opus 5.5 $10.04, Astra $6.24, Sol $1.27.
  • No model leads all nine: Opus 5.5 leads three, including FrontierCode 2.0 at 46%; Astra leads Terminal-Bench-Science (63%) and CritPt v2 (72%).
  • Median output tokens per task rose from 35k in v4.3 to 59k.

What it means for you

Effort settings now move scores a lot: Opus 5.5 ranges from 35 at low effort to 58 at max, at ten times the cost per task. When you compare models, check which effort a score refers to and what it costs to reach it.

The index is mainly text-based and English-language, according to the methodology page, so image, speech and multilingual work need separate evidence.

Sources & further reading

Based on the official documentation below. Features and access can change; check the provider before making a decision.

  1. Artificial Analysis: Announcing Intelligence Index v5 (opens in a new tab)
  2. Artificial Analysis: Intelligence benchmarking methodology (opens in a new tab)
Something out of date or incorrect?Report a correction

Was this page useful?

Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.

Share this page

Related reading