reasoning benchmark
MuSR
MuSR (Multistep Soft Reasoning) is a benchmark for evaluating language models on multistep soft reasoning tasks specified in natural language narratives. Created through a neurosymbolic synthetic-to-natural generation algorithm, it generates complex reasoning scenarios like murder mysteries roughly 1000 words in length that challenge current LLMs including GPT-4. The benchmark tests chain-of-thought reasoning capabilities across domains involving commonsense reasoning about physical and social situations.
Updated Aug 11, 2026
Models2
Model coverage2
MetricScore
EvidenceB
MuSR Ranking
Higher score ranks better on this benchmark.
| 01 | MA | 76.4% | 100.0% | 2 | C | |
| 02 | NR | 50.7% | 0.0% | 2 | C |
MuSR Score Distribution
A closer view of the leading scores on this benchmark.
MuSR Highlights
The leading models and scores on this benchmark.
What is MuSR?
What MuSR measures and how its scores work.
MuSR (Multistep Soft Reasoning) is a benchmark for evaluating language models on multistep soft reasoning tasks specified in natural language narratives. Created through a neurosymbolic synthetic-to-natural generation algorithm, it generates complex reasoning scenarios like murder mysteries roughly 1000 words in length that challenge current LLMs including GPT-4. The benchmark tests chain-of-thought reasoning capabilities across domains involving commonsense reasoning about physical and social situations.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MuSR
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- musr|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MuSR.
Which model scores highest on MuSR?
Kimi K2 Instruct is currently ranked first with 76.4%.
What does MuSR measure?
MuSR (Multistep Soft Reasoning) is a benchmark for evaluating language models on multistep soft reasoning tasks specified in natural language narratives. Created through a neurosymbolic synthetic-to-natural generation algorithm, it generates complex reasoning scenarios like murder mysteries roughly 1000 words in length that challenge current LLMs including GPT-4. The benchmark tests chain-of-thought reasoning capabilities across domains involving commonsense reasoning about physical and social situations.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
2 model results are currently shown.
Does this benchmark affect the overall score?
No. This benchmark is shown for reference but does not contribute to the overall score.