AIEvals

Benchmarks

FrontierMath Tier 4 (v2)

Research-level mathematics problems, with Tier 4 the hardest slice of the set.

MathematicsEpoch AIVendor-reported onlyReported by OpenAI

How it is run

OpenAI reports Tier 4 only, where Astra scores 97.6% against 83.0% for GPT-5.6 Sol.

What it cannot tell you

Only the hardest tier is published here, so this row says nothing about performance across the rest of the set. OpenAI reports that Astra has saturated the tier.

Why there is no independent table

No independent suite in this dataset runs FrontierMath Tier 4 (v2). The numbers below are the ones model publishers report about their own models, which is a different kind of evidence: each publisher chooses its own harness and effort level, so a row is readable on its own and not against the other rows.

Read this with

Benchmarks in the same category, which usually disagree with this one on purpose.

  • AIME 2025Same category. Olympiad mathematics from the 2025 American Invitational Mathematics Examination.
  • AIME 2026Same category. Competition mathematics from the 2026 American Invitational Mathematics Examination.
  • RiemannBenchSame category. Mathematical reasoning, included in the report’s science and math evaluations.
  • ProofBench v1.1Same category. Formal and semi-formal mathematical proof construction, graded on the proof rather than the final answer.
  • MathArena ApexSame category. Final-answer mathematics problems that no model had solved when the set was published — Apex exists because the standard competitions had stopped discriminating.

As reported by each publisher

Kept separate per publisher, because the settings are not shared.

OpenAI

GPT-6 Astra: evaluation table · 3 September 2026

RankModelScore
1GPT-6 Astra97.6%
4GPT-5.6 Sol83.0%
3Claude Fable 5.187.8%
2Claude Fable 590.2%
5Claude Opus 573.2%

Ranks compare reported scores in the same published row; ties share a rank.

OpenAI reports the maximum score at any effort level, in its own research environment or through its API. Claude and Gemini figures are taken by OpenAI from those vendors' published reports, so this table is a comparison between publication claims, not between runs.

openai.com/index/gpt-6-astra/

Sources

Every page used for this benchmark, read on 17 October 2026.