OptiVerse - Mathematical Programming: leaderboard

OptiVerse (2026) is a benchmark of textbook-sourced optimization problems that models solve by writing and running solver code, a problem counting as solved only when every required value matches the ground truth within 0.1%. This board is accuracy in percent on the mathematical programming problems; higher is better.

Metric: Accuracy (%) on the 367 mathematical programming problems of OptiVerse; the model writes and runs solver code (gurobi, pyomo, cvxpy, ortools and others) and a problem counts as solved only when every required variable and the objective match the ground truth within 0.1% relative error, extracted and checked by a DeepSeek-V3.2-Chat judge; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro58.04#145
2GPT-5.2Thinking55.86#185
3Gemini 3 Flash54.77#175
4Gemini 2.5 Pro53.68#276
5DeepSeek V3.2Thinking53.68#419
6Qwen3-235B-Thinking (OptiVerse checkpoint unspecified)53.68
7Claude Sonnet 4.5Thinking53.41#261
8O352.04#237
9O4 Mini50.95#368
10Qwen3-30B-Thinking (OptiVerse checkpoint unspecified)49.59
11Gemini 2.5 Flash46.05#463
12Qwen 3 235B Instruct (OptiVerse checkpoint unspecified)44.41
13Qwen 3 8BThinking40.33#1873
14Kimi K240.05#443
15Qwen 3 30B Instruct (OptiVerse checkpoint unspecified)38.96

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

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