Skip to main content
Evaluation sources/MysteryMechanism
ParticipatingReasoning

MysteryMechanism

Vals AI · Public model evaluation results from MysteryMechanism.

Official evaluation
Source description

This source measures concrete delivery performance in Reasoning; raw scores are shown separately from the normalized capability score.

Limitation: Different versions, reasoning settings, and test sets may change scores; read them with the data date.

How it is used

Participating sources affect capability scores; reference, cross-reference, and observing sources are for comparison and verification only.

Model results from this source

GPT-6 Astra53.1574.32026-11-05
Claude Fable 5.251.872.62026-11-10
Claude Opus 5.549.5569.92026-11-05
Claude Sonnet 5.549.169.32026-11-05
Claude Fable 5.147.7567.72026-11-05
GPT-6.1 Sol46.466.02026-11-05
Gemini 4 Argon45.4964.92026-11-05
Claude Opus 537.3955.02026-11-05
Gemini 3.8 Flash36.4954.02026-11-05
Muse Spark 1.336.0453.42026-11-05
Claude Haiku 5.535.1352.32026-11-05
GPT-5.6 Sol33.3350.12026-11-05
Grok 4.630.6346.82026-11-05
GPT-6 Sol30.1846.22026-11-05
Grok 4.725.2340.22026-11-05
Qwen3.8 Max23.8738.52026-11-05
GLM 5.322.9737.42026-11-05
MiMo V2.6 Flash21.6235.82026-11-05
MiMo V2.6 Flash21.6235.82026-10-17
MiMo V2.6 Flash21.6235.82026-10-18
DeepSeek V4.1 Flash21.1735.22026-11-05
GPT-6 Luna19.3733.02026-11-05
Step 5 Preview18.4732.02026-11-05
MiMo V2.6 Pro15.3128.12026-11-05
GPT-5.6 Luna14.4127.02026-11-05
Mistral Large 412.1624.22026-10-31

Frequently asked questions