Is FrontierMath Erdős still meaningful?
Status:Still useful Checked against the report, 5 November 2026
Verdict
Yes. It still leaves almost all of its room at the top, though the set is small.
Why still useful: The best model solves 4.4% (Sep 2026), so models still have clear room to climb.[1][1] Benchmark data bundle Epoch AI Primary
What “Still useful” meansKnown problems
- The set is small, so one solved problem moves the score by 1.5 points.[1][1] Benchmark data bundle Epoch AI Primary
What to look at instead
The report names no direct successor.
Where it appears in the report
Chapter 1, Benchmarks
- Figure 1. Each New Benchmark Fills Up Faster Than the Last
- What still resists
- Figure 4. Best score on benchmarks that still resist, November 2026
- Table 5. Reasoning and knowledge benchmarks: what they measure, who maintains them, the top score on 5 November 2026 and the known flaw in “Go deeper: Reasoning and knowledge benchmarks (17)”
Sources
Copied from the chapter source lists. Each number links back to the source in its chapter.
- [1] Benchmark data bundle, Epoch AI. Dated scores (5 Nov 2026) and ceilings for most benchmarks here. Primary epoch.ai Chapter 1, source 1
- [2] FrontierMath, Epoch AI. It describes the tiers, Open Problems and Erdős sets. Primary epoch.ai Chapter 1, source 25
- [3] Benchmarking Hub, Epoch AI. It explains the hub, the ECI and which benchmarks Epoch runs itself. Primary epoch.ai Chapter 1, source 26