MM-OptBench: A Solver-Grounded Benchmark for Multimodal Optimization Modeling: leaderboard
MM-OptBench evaluates multimodal optimization modeling from text-and-visual problem specifications, where models must produce a mathematical formulation and executable solver code. The pinned metric, pass@1 (%), measures the share of instances solved on the first attempt in the solver-grounded evaluation. Source: BenchmarkList’s attributed result table.
Metric: pass@1 (%). Source: benchmarklist.com. Saturation forecast: Around February 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 52.1 | #83 |
| 2 | Gemini 3.1 Pro (Preview) | 51.3 | #64 |
| 3 | Claude Sonnet 4.6 | 23.5 | #94 |
| 4 | Qwen-VL-Max | 16.8 | #475 |
| 5 | GLM-4.5V | 15.4 | #358 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mm-optbench-a-solver-grounded-benchmark-for-multimodal-optimization-modeling · How It Works · Data refreshed daily, snapshot 2026-11-05.