OptiVerse - Easy: 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 easy problems; higher is better.

Metric: Accuracy (%) on the 300 easy-level 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. 20 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.2Thinking91#185
2Gemini 3 Pro89#145
3Gemini 3 Flash88.67#175
4O4 Mini87.67#368
5Qwen3-235B-Thinking (OptiVerse checkpoint unspecified)87.33
6Gemini 2.5 Pro87#276
7O386.67#237
8DeepSeek V3.2Thinking84.33#419
9Claude Sonnet 4.5Thinking83.67#261
10Gemini 2.5 Flash82.33#463
11Qwen3-30B-Thinking (OptiVerse checkpoint unspecified)80.67
12GPT-OSS-120B78.67#704
13Qwen 3 235B Instruct (OptiVerse checkpoint unspecified)78.33
14Qwen 3 8BThinking73#1873
15Kimi K271.33#443

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

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