As of November 9, 2026, Polymarket prices “Will the highest score achieved by an Anthropic Claude model on Humanity’s Last Exam in 2026 be 60% or higher?” at 91% YES with $75K traded. No tracked wallet holds a position on this market, so there is no verdict.
Claude Fable 5.1 is listed at 65.0% on Humanity's Last Exam with tools for a September 2026 release, which would clear the 60% threshold if that figure holds.
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026, with Fable 5.1 posting a 65.0% score on Humanity's Last Exam with tools and 60.9% without tools, according to the company's launch materials. The result marked a substantial jump over its predecessor, Claude Fable 5, which Anthropic said scored lower on the benchmark. The figure is directly relevant to the question of whether the highest score achieved by an Anthropic Claude model on Humanity's Last Exam in 2026 will be 60% or higher, since it cleared that threshold on the first day of September. [Anthropic, Sep 01]
Subsequent releases pushed Anthropic's reported numbers higher still. Claude Opus 5.5, launched September 22, 2026, was credited by Anthropic with 67.7% on Humanity's Last Exam with tools, though independent evaluator Artificial Analysis measured 61.4% — a gap of 6.3 points. Days later, Claude Sonnet 5.5 was reported at 64.5% with tools, up from Sonnet 5's 54.9% and close to Opus 5.5's figure. VentureBeat noted that Opus 5.5 delivered some of the highest benchmark results yet from any Anthropic model across agentic coding, knowledge work and multidisciplinary reasoning. [VentureBeat, Sep 22]
The divergence between vendor-reported and independent scores is the central uncertainty for the highest score achieved by an Anthropic Claude model on Humanity's Last Exam in 2026 reaching 60% or higher. Anthropic's own figures for Fable 5.1, Opus 5.5 and Sonnet 5.5 all exceed the threshold, while Artificial Analysis' 61.4% for Opus 5.5 clears it by a narrower margin. With Anthropic's release cadence running roughly every three weeks through the fall, additional models could be introduced before year-end, leaving room for both further gains and further measurement disputes. [Kingy AI, Sep 22]
None of the 166 tracked wallets holds a position here, so the Radar makes no call — price, volume and news above are live. When a tracked wallet takes a position, this page gets a verdict.
See the markets with a verdict →