Skip to content

Benchmark

WildClawBench · Pure-text (34 tasks)

Scores are percentages. Higher is better.

Models and systems measured
34
Scores recorded
34
Sources
1
IQ weight
1.81

Score distribution

49.4%
Mean
51.7%
Median
9.8%
Std dev
36%
10th percentile
62.7%
90th percentile
28.1%
Min
28.1%65.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.81 for chat models, measured over 15 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
1GPT-5.6 SolOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs65.6%
2Claude Opus 4.7OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs65%
3Claude Opus 4.8OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs64.3%
4Kimi K3OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs62.7%
5Grok 4.5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs59.2%
6GPT 5.4OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs58%
7Claude Fable 5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs57%
8GLM 5.2OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs56.7%
9Muse Spark 1.1OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs55.7%
10Hy3 (Deep Infra)OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs54.8%
11Qwen 3.8 27BOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs54.8%
12Claude Opus 4.6OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs54.6%
13GPT 5.5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs54.5%
14GLM 5.1OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs53.9%
15Qwen3.8 MaxOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs53.7%
16GLM 5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs51.7%
17DeepSeek V4 Pro 0813OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs51.4%
18Kimi K2.7 CodeOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs50.9%
19Muse Glimmer 30BOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs48.4%
20MiMo V2 ProOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs48.2%
21MiMo V2.5 ProOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs48.1%
22Qwen3.6 27BOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs47%
23MiniMax M2.7OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs44.9%
24Qwen3.5 397B A17BOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs42.6%
25DeepSeek V3.2OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs40.1%
26Gemma 4 31B ITOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs39.8%
27Gemini 3.1 Pro PreviewOpenClaw · pure-text split · 34 tasks · low-effort· Reported by WildClawBench / InternLM· via BenchBoard docs38.7%
28Step 3.5 FlashOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs37.4%
29Kimi K2.5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs36%
30MiniMax M2.5OpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs35.3%
31MiMo V2 FlashOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs28.1%

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
GLM 5 TurboOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs41%
Grok 4.20 BetaOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs28.4%
Intern-S2 Preview 397BOpenClaw · pure-text split · 34 tasks· Reported by WildClawBench / InternLM· via BenchBoard docs52.7%

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.