Rapid progress on Epoch AI’s FrontierMath benchmark, with leading models like GPT-5.6 Sol and Claude Fable 5 reaching 87-89% on Tier 4 (v2) and legacy versions as of mid-2026, underpins the 90% market-implied odds for crossing 90% before 2027. Released in late 2024 with initial top scores below 2%, the benchmark saw scores climb quickly after OpenAI’s o3 demonstrated ~25-30% capability, followed by iterative gains through GPT-5 iterations and Anthropic releases. The June 2026 v2 update refined 42% of problems, yet frontier systems continued closing the gap via improved reasoning, tool use, and training scale. Traders anticipate further model launches and refinements in the remaining months of 2026 will push at least one system over the threshold, consistent with the observed acceleration in large language model mathematical performance.
Experimental AI-generated summary referencing Polymarket data. This is not trading advice and plays no role in how this market resolves. · Updated$117,315 Vol.
$117,315 Vol.
$117,315 Vol.
$117,315 Vol.
The primary resolution source will be information from EpochAI however a consensus of credible reporting may also be used.
Market Opened: Nov 12, 2025, 5:15 PM ET
Resolver
0x65070BE91...The primary resolution source will be information from EpochAI however a consensus of credible reporting may also be used.
Resolver
0x65070BE91...Rapid progress on Epoch AI’s FrontierMath benchmark, with leading models like GPT-5.6 Sol and Claude Fable 5 reaching 87-89% on Tier 4 (v2) and legacy versions as of mid-2026, underpins the 90% market-implied odds for crossing 90% before 2027. Released in late 2024 with initial top scores below 2%, the benchmark saw scores climb quickly after OpenAI’s o3 demonstrated ~25-30% capability, followed by iterative gains through GPT-5 iterations and Anthropic releases. The June 2026 v2 update refined 42% of problems, yet frontier systems continued closing the gap via improved reasoning, tool use, and training scale. Traders anticipate further model launches and refinements in the remaining months of 2026 will push at least one system over the threshold, consistent with the observed acceleration in large language model mathematical performance.
Experimental AI-generated summary referencing Polymarket data. This is not trading advice and plays no role in how this market resolves. · Updated



Beware of external links.
Beware of external links.
Frequently Asked Questions