BlogLLM Benchmark Scores as Published, Side by Side
PerceptionBench: Results as Published
Every result the catalog holds for PerceptionBench: 10 models from 4 developers, read from 2 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Multimodal UnderstandingMaintainer Moonshot AIIntroduced 2026Data version 2026103001, as of 2026-10-30
PerceptionBench is an in-house benchmark of Moonshot AI that focuses on atomic visual perception capabilities, the low-level reading of what is in an image rather than reasoning about it. The card prints one number per model, averaged over three runs, and does not name the metric or the item count.
PerceptionBench
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 77.6% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Tools: none; Fallback: with fallback |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 59.7% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: max |
| GPT-5.5OpenAI, released 2026-04-23 | 55.8% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: xhigh |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 72.1% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Tools: none; Fallback: with fallback |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 63% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Tools: none; Fallback: with fallback |
| Claude Fable 5Anthropic, released 2026-06-09 | 57.2% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: max; Fallback: w/ fallback |
| Claude Opus 4.8Anthropic, released 2026-05-28 | 47.2% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: max |
| DeepSeek-V4.1-Flash-1017DeepSeek, released 2026-10-17 | 47.7% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Tools: none |
| Kimi K3.1 ProMoonshot AI, released 2026-10-23 | 68.3% | Moonshot AI, Introducing Kimi K3.1 Prochecked 2026-10-24 | Effort: max; Tools: none |
| Kimi K3Moonshot AI, released 2026-07-16 | 58.5% | Moonshot AI, Kimi K3 model cardchecked 2026-10-15 | Effort: max |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.