# Artificial Analysis releases Intelligence Index v5 with private datasets

Digest AI · Research · published 2026-10-28T03:06:22Z

Canonical: https://digestai.news/story/artificial-analysis-releases-intelligence-index-v5-with-private-datase

## Summary

Artificial Analysis has launched version 5 of its Intelligence Index, a composite measure of AI model capabilities. This update marks the first time the majority of the index relies on private datasets, with 60% of evaluations now using private test sets or held-out answers to prevent contamination. The new index introduces two benchmarks: FrontierCode 2.0, a private coding test from Cognition, and Terminal-Bench-Science, which assesses agentic scientific research tasks. Existing benchmarks like HLE-Diamond and AA-Omniscience have also been upgraded.

Claude Opus 5.5 (max) leads the new index with a score of 58, narrowly ahead of GPT-6 Astra (max) at 57 and GPT-6.1 Sol (max) at 55. Anthropic and OpenAI models occupy the top five positions. The gap between closed and open-weight models has widened significantly; the best open-weight model, Kimi K3.1 Pro, scores 46, trailing the leader by 12 points compared to a 3-point gap in the previous version. Among released open weights, DeepSeek V4.1 Flash 1017 and GLM-5.3 lead at 32.

The update highlights substantial cost variations, with GPT-6.1 Terra being the cheapest at $0.96 per task and Claude Fable 5.1 the most expensive at $13.04. Artificial Analysis also introduced the Progress Index v1, which tracks model capability over time on a continuous scale for models released since November 2022.

## Key points

- Intelligence Index v5 uses 60% private datasets to prevent benchmark contamination.
- Claude Opus 5.5 leads with a score of 58, ahead of GPT-6 Astra at 57.
- Open-weight models trail the leader by 12 points, a gap that widened from 3 points.

## Why it matters

The shift to private benchmarks reduces data contamination, offering a more accurate view of real-world AI capabilities. The widening gap between closed and open models signals a significant divergence in performance and cost efficiency for enterprise adoption.

## Sources

1. [Artificial Analysis Intelligence Index v5](https://artificialanalysis.ai/articles/artificial-analysis-intelligence-index-v5) (artificialanalysis.ai, 2026-10-28)

Part of the developing story: [AI Model Wars From Replit to Claude Opus](https://digestai.news/thread/replit-ends-partnership-with-anthropic-cites-cost-and-performance) (5 stories)

## Cite

Digest AI, "Artificial Analysis releases Intelligence Index v5 with private datasets", 28 October 2026, https://digestai.news/story/artificial-analysis-releases-intelligence-index-v5-with-private-datase

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/artificial-analysis-releases-intelligence-index-v5-with-private-datase.json