Skip to content

Benchmark

PinchBench v2 · Best Success Rate

Scores are percentages. Higher is better.

Models and systems measured
60
Scores recorded
64
Sources
1
IQ weight
1.39

Score distribution

79.3%
Mean
85.3%
Median
18.2%
Std dev
58.9%
10th percentile
91.9%
90th percentile
3.2%
Min
3.2%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.39 for chat models, measured over 50 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
51GPT OSS 20bOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs41.8%
52GPT-5.5 ProOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs37.1%
53Llama 3.1 70B InstructOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs10.7%
54Llama 4 ScoutOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs3.2%

Showing 51–54 of 54 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Aion 3.0OpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs62.4%
GLM 5 TurboOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs88.2%
GLM 5V TurboOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs86.6%
Grok 4.20OpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs87.8%
Grok 4.3OpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs85%
Grok Build 0.1OpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs92.1%
Ling 2.6 1TOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs82.6%
Nemotron 3.5 Bolt 74B-A9B 24GBOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs90.1%
Nemotron 3.5 Bolt 74B-A9B BF16OpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs91.7%
Seed-2.0-LiteOpenClaw · PinchBench v2.0.0 · task count: 147 on live leaderboard / 148 in v2.0.0 release notes · best run· Reported by PinchBench· via BenchBoard docs89.7%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.