Skip to content

Benchmark

QwenClawBench v1.1 · Final Score

Scores are percentages. Higher is better.

Models and systems measured
14
Scores recorded
16
Sources
1
IQ weight
1.32

Score distribution

58.2%
Mean
57.4%
Median
5.7%
Std dev
51.9%
10th percentile
68.5%
90th percentile
50.5%
Min
50.5%70.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.32 for chat models, measured over 13 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
1Qwen3.7 MaxQwen3.7 release cross-harness chart · Hermes· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.70.7%
2Qwen3.7 MaxQwen3.7 release cross-harness chart · Claude Code· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.68.5%
3Qwen3.7 MaxQwen3.7-Max release · OpenClaw · release table and cross-harness figure· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.64.3%
4Qwen3.7 PlusQwen3.7-Plus release · release table · harness not separately disclosed· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.61.8%
5Claude Sonnet 4.6OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs60.5%
6Claude Opus 4.6OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs59.5%
7GLM 5.1OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs58.7%
8Qwen3.6 PlusOpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs57.4%
9GPT 5.4OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task · high effort· Reported by QwenClawBench / Qwen Team· via BenchBoard docs56.7%
10MiMo V2 ProOpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs56.5%
11Claude Opus 4.5OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task · high effort· Reported by QwenClawBench / Qwen Team· via BenchBoard docs56.1%
12GLM 5OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs54.9%
13Kimi K2.5OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs51.9%
14DeepSeek V3.2Thinking mode · OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs50.7%
15MiniMax M2.7OpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task · high effort· Reported by QwenClawBench / Qwen Team· via BenchBoard docs50.5%

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
Qwen3.5 PlusOpenClaw · qwenclawbench-v1.1-100 · 100 tasks · 3 runs/task· Reported by QwenClawBench / Qwen Team· via BenchBoard docs53%

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.