MMAnswerBench
We show this table for reference; we do not rank on it.
A multimodal mathematical reasoning benchmark that tests whether models can answer visually grounded math questions correctly.
Benchmark score on MMAnswerBench — November 5, 2026
We compile the MMAnswerBench rows from provider self-reports and secondary reports. GLM-5.2 leads the table at 91.0%, followed by Kimi K2.6 (86.0%) and Claude Opus 4.5 (84.0%). We do not use these results to rank models overall.
GLM-5.2
Z.AI
Kimi K2.6
Moonshot AI
Claude Opus 4.5
Anthropic
10 modelsMathematicsCurrentDisplay onlyUpdated November 5, 2026
Benchmark score table (10 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | GLM-5.2Z.AI | 91.0% | Not reported | Open |
| 2 | Kimi K2.6Moonshot AI | 86.0% | Not reported | Open |
| 3 | Claude Opus 4.5Anthropic | 84.0% | Not reported | Closed |
| 4 | Qwen3.6 PlusAlibaba | 83.8% | Not reported | Closed |
| 5 | GLM-5.1Z.AI | 83.8% | Not reported | Open |
| 6 | GLM-5Z.AI | 82.5% | Not reported | Open |
| 7 | Kimi K2.5Moonshot AI | 81.8% | Not reported | Open |
| 8 | Qwen3.5 397BAlibaba | 80.9% | Not reported | Open |
| 9 | Qwen3.6-27BAlibaba | 80.8% | Not reported | Open |
| 10 | Qwen3.6-35B-A3BAlibaba | 78.9% | Not reported | Open |
Among the reported MMAnswerBench rows, GLM-5.2 is first at 91.0%. The third row is 7.0 points behind. The broader top-10 range is 12.1 points, so the table still separates the published systems.
10 models have been evaluated on MMAnswerBench. The benchmark falls in the Mathematics category. MMAnswerBench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About MMAnswerBench
Year
2026
Tasks
Multimodal math questions
Format
Visual and structured mathematical QA
Difficulty
Advanced mathematical reasoning
MMAnswerBench matters because text-only math ability does not guarantee strong performance when the relevant information is embedded in diagrams, tables, or other visual inputs. It acts as a multimodal math transfer check.
Freshness and provenance
Version
MMAnswerBench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does MMAnswerBench measure?
A multimodal mathematical reasoning benchmark that tests whether models can answer visually grounded math questions correctly.
Which model scores highest on MMAnswerBench?
GLM-5.2 by Z.AI currently leads with a score of 91.0% on MMAnswerBench.
How many models are evaluated on MMAnswerBench?
10 AI models have published results on MMAnswerBench in the BenchLM catalog.
Compare top models on MMAnswerBench
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.