Benchmark
PerceptionBench
PerceptionBench is Moonshot AI's internal benchmark for evaluating atomic visual perception capabilities.
- Modality
- multimodal
- Categories
- multimodal, reasoning, vision
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 2
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Kimi K3 | Moonshot AI | 0.585 | 2026-07-16 | evidence |
| Qwen3.8 Max | Qwen | 0.635 | 2026-08-02 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.