Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Multimodal & Grounded benchmark
Every model the catalog carries a published We-Math value for, ranked by that value.
A multimodal math benchmark for visually grounded mathematical reasoning and answer generation. 6 models in the ModelScale catalog have a published value.
Each point is a model with a published We-Math value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
6 models with a published We-Math 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 | Multimodal mathematical reasoning |
|---|---|---|---|
| 1 | Alibaba | 87.9 | |
| 2 | 86.9 | ||
| 3 | Moonshot AI | 84.7 |
What the benchmark measures and how its value is published, from the source’s own record.
We-Math is a multimodal & grounded benchmark published by Qwen3.6 launch benchmarks. A multimodal math benchmark for visually grounded mathematical reasoning and answer generation. Its evaluation set is described by the source as visually grounded math problems. The source rates its difficulty as advanced multimodal mathematics.
Each model's result is published by Qwen3.6 launch benchmarks as multimodal mathematical reasoning. ModelScale prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 6 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
All leaderboards · Multimodal & Grounded capability leaderboard
| 4 | OpenAI | 79.0 |
| 5 | Anthropic | 70.0 |
| 6 | LiquidAI | 68.6 |
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.