Artificial Analysis released version 4.2 of its Intelligence Index on September 5, 2026, saying it aimed to address skepticism about its benchmarks failing to reflect GPT-6 Astra's true capabilities. It is the company's first major update to the index since its initial release.
Artificial Analysis reported GPT-6 Astra scored 169 points across over 50 benchmarks, placing it first among 267 models, according to Epoch AI. That compares with a score that was previously on par with its predecessor.
The updated index includes two new benchmarks: AA-Briefcase for real-world knowledge work and GDP.pdf from Surge AI for PDF document analysis. GPQA-Diamond has been dropped because models have solved it.
"Artificial Analysis held off on updates to keep scores stable during major model launches, but the top of the leaderboard moved so fast that an interim update was necessary," said Matthias Bastian, the source of the report. The company also fixed scoring errors across several benchmarks and tweaked its grading systems for more stable results.
The announcement follows criticism that earlier benchmarks failed to capture GPT-6 Astra's actual progress. Artificial Analysis said it is working on version 5, which has been in development for eight months and will roll out in stages.
Artificial Analysis did not say when version 5 will be fully released, and it raised questions about how to measure progress in a rapidly evolving field. The company said it will continue refining its benchmarks as models advance.
Source: thedecoder
