Skip to content

Benchmark

Galileo Agent Leaderboard

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.

Models measured
21
Scores recorded
126
Sources
1
IQ weight
1.15

Score distribution

40.7%
Mean
42%
Median
14.8%
Std dev
18%
10th percentile
60%
90th percentile
11%
Min
11%74%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.15 for chat models, measured over 21 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen3 235B A22B Instruct 2507Insurance· Galileo Agent Leaderboard data74%
2Mistral Medium 3.1Insurance· Galileo Agent Leaderboard data70%
3GPT 4.1Insurance· Galileo Agent Leaderboard data66%
4GPT 4.1Investment· Galileo Agent Leaderboard data64%
5GPT-4.1 miniTelecom· Galileo Agent Leaderboard data64%
6Claude Sonnet 4Healthcare· Galileo Agent Leaderboard data62%
7GPT 4.1· Galileo Agent Leaderboard data62%
8GPT 4.1Healthcare· Galileo Agent Leaderboard data62%
9Mistral Medium 3.1· Galileo Agent Leaderboard data61%
10Qwen2.5 72B InstructHealthcare· Galileo Agent Leaderboard data61%
11Gemini 2.5 Flash-LiteHealthcare· Galileo Agent Leaderboard data60%
12GPT 4.1Banking· Galileo Agent Leaderboard data60%
13GPT-4.1 miniHealthcare· Galileo Agent Leaderboard data60%
14Mistral Medium 3.1Healthcare· Galileo Agent Leaderboard data60%
15Mistral Medium 3.1Telecom· Galileo Agent Leaderboard data59%
16Claude Sonnet 4Banking· Galileo Agent Leaderboard data58%
17GPT 4.1Telecom· Galileo Agent Leaderboard data58%
18Kimi K2 InstructBanking· Galileo Agent Leaderboard data58%
19Kimi K2 InstructInsurance· Galileo Agent Leaderboard data58%
20Qwen3 235B A22B Instruct 2507Telecom· Galileo Agent Leaderboard data58%
21Mistral Medium 3.1Banking· Galileo Agent Leaderboard data57%
22Mistral Medium 3.1Investment· Galileo Agent Leaderboard data57%
23GPT-4.1 mini· Galileo Agent Leaderboard data56%
24GPT-4.1 miniBanking· Galileo Agent Leaderboard data56%
25Claude Sonnet 4· Galileo Agent Leaderboard data55%
26Gemini 2.5 Flash-LiteInsurance· Galileo Agent Leaderboard data54%
27Gemini 2.5 ProInsurance· Galileo Agent Leaderboard data54%
28Claude Sonnet 4Insurance· Galileo Agent Leaderboard data53%
29Claude Sonnet 4Telecom· Galileo Agent Leaderboard data53%
30GLM 4.5 AirInsurance· Galileo Agent Leaderboard data53%
31Kimi K2 Instruct· Galileo Agent Leaderboard data53%
32Kimi K2 InstructTelecom· Galileo Agent Leaderboard data53%
33Qwen3 235B A22B Instruct 2507· Galileo Agent Leaderboard data53%
34Qwen2.5 72B InstructInsurance· Galileo Agent Leaderboard data52%
35Qwen2.5 72B InstructTelecom· Galileo Agent Leaderboard data52%
36Qwen2.5 72B Instruct· Galileo Agent Leaderboard data51%
37GPT-4.1 miniInvestment· Galileo Agent Leaderboard data50%
38Grok 4Investment· Galileo Agent Leaderboard data50%
39Claude Sonnet 4Investment· Galileo Agent Leaderboard data49%
40GLM 4.5 AirBanking· Galileo Agent Leaderboard data49%
41Kimi K2 InstructHealthcare· Galileo Agent Leaderboard data49%
42Qwen3 235B A22B Instruct 2507Healthcare· Galileo Agent Leaderboard data49%
43DeepSeek V3Insurance· Galileo Agent Leaderboard data48%
44Gemini 2.5 FlashBanking· Galileo Agent Leaderboard data48%
45Grok 4Insurance· Galileo Agent Leaderboard data48%
46Qwen2.5 72B InstructBanking· Galileo Agent Leaderboard data48%
47DeepSeek V3Telecom· Galileo Agent Leaderboard data47%
48Gemini 2.5 Flash-Lite· Galileo Agent Leaderboard data47%
49Kimi K2 InstructInvestment· Galileo Agent Leaderboard data47%
50GLM 4.5 AirInvestment· Galileo Agent Leaderboard data46%

Showing 1–50 of 126 results.

Data credits: Galileo Agent Leaderboard, Apache-2.0 (licence)