Artificial Analysis Intelligence Index Upgraded to v5

Artificial Analysis released v5 of its Intelligence Index, calling it more accurate, but the gap between open- and closed-weight models widened from 3 to 12 points. With 60% of the index based on private test sets or held-out answers, results cannot be reproduced, and some users already question its accuracy.

• Jev calibration benchmark: LLMs beat humans yet know when they're right2026-09-22
• Defense of Artificial Analysis Benchmarks2026-09-11
• Cloudflare discloses cross-request KV cache leak in GLM-5.32026-10-27
• Moonshot resets subscriber quotas after Kimi K3.1 Pro launch issues2026-10-26
• Moonshot AI releases Kimi K3.1 Pro model upgrade2026-10-23