Mathematics benchmark
MMAnswerBench leaderboard
Every model the catalog carries a published MMAnswerBench value for, ranked by that value.
CategoryMathematics
MeasureVisual and structured mathematical QA
TasksMultimodal math questions
DifficultyAdvanced mathematical reasoning
Published byQwen3.6 launch benchmarks
A multimodal mathematical reasoning benchmark that tests whether models can answer visually grounded math questions correctly.
MMAnswerBench ranking
10 models with a published MMAnswerBench value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Visual and structured mathematical QA |
|---|---|---|---|
| 1 |  GLM-5.2 | Z.AI | 91.0 |
| 2 |  Kimi K2.6 | Moonshot AI | 86.0 |
| 3 |  Claude Opus 4.5 | Anthropic | 84.0 |
| 4 | GLM-5.1 | Z.AI | 83.8 |
| 4 |  Qwen3.6 Plus | Alibaba | 83.8 |
| 6 | GLM-5 | Z.AI | 82.5 |
| 7 | Kimi K2.5 | Moonshot AI | 81.8 |
| 8 | Qwen3.5 397B | Alibaba | 80.9 |
| 9 | Qwen3.6-27B | Alibaba | 80.8 |
| 10 | Qwen3.6-35B-A3B | Alibaba | 78.9 |
Evidence key: Observed
Rows are ordered by the value Qwen3.6 launch benchmarks published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.