MathDuels: leaderboard

MathDuels (2026) is a self-play mathematics benchmark: each model authors 30 hardened problems and solves those the other models wrote, with symbolic answer checking and a verifier removing ill-posed problems; ratings come from a Rasch model on an Elo-style scale. This board is the composite rating, the mean of the solver and author ratings, in rating points on an open scale; higher is better.

Metric: Composite rating (Elo-style points, open scale): the mean of the solver and author ratings; 19 models each author 30 hardened problems (six mathematical domains) and solve every problem authored by the others; answers checked symbolically, ill-posed problems removed by a GPT-5.4-high verifier; Rasch model on a shared rating scale anchored at Grok-4.1-fast-high = 1500, k = 30; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3.1 Pro (Preview)High1919#99
2GPT-5.4High1870#135
3GPT-5.2High1737#185
4Claude Opus 4.6High1675#113
5Gemini 3 FlashHigh1673#175
6Claude Sonnet 4.6High1593#173
7GPT-5.4 MiniHigh1570#406
8Qwen 3.5 397B A17B1548#264
9Kimi K2.51542#266
10Grok 4.20High1485#301
11DeepSeek V3.21449#419
12GLM-51438#269
13MiniMax-M2.71372#539
14Grok 4.1 FastHigh1335#427
15Step 3.5 Flash1328#526

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=mathduels · How It Works · Data refreshed daily, snapshot 2026-11-21.