Skip to content

Benchmark

Kagi LLM Benchmark

Scores are percentages.

Models measured
110
Scores recorded
123
Sources
1
IQ weight
1.63

Score distribution

54.3%
Mean
55.2%
Median
17.8%
Std dev
31.5%
10th percentile
75%
90th percentile
6.3%
Min
6.3%91.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.63 for chat models, measured over 110 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model
1Claude Fable 5thinking· Kagi LLM Benchmark91.4%
2Claude Fable 5· Kagi LLM Benchmark88.8%
3Claude Opus 4.8thinking· Kagi LLM Benchmark88.8%
4GPT 5.5· Kagi LLM Benchmark88.8%
5Claude Opus 4.6thinking· Kagi LLM Benchmark83.6%
6Grok 4.5· Kagi LLM Benchmark83.5%
7Claude Opus 4.7thinking· Kagi LLM Benchmark80.7%
8Claude Opus 4.5thinking· Kagi LLM Benchmark80.2%
9Gemini 3 Pro Preview· Kagi LLM Benchmark80.1%
10Kimi K2.5thinking· Kagi LLM Benchmark78.5%
11GPT-5 Pro· Kagi LLM Benchmark76.8%
12GLM 5thinking· Kagi LLM Benchmark75%
13Grok 4.20thinking· Kagi LLM Benchmark75%
14Claude Opus 4thinking· Kagi LLM Benchmark74.3%
15Qwen3.5 397B A17B· Kagi LLM Benchmark73.7%
16Grok 4· Kagi LLM Benchmark73.6%
17Claude Opus 4.7· Kagi LLM Benchmark73.3%
18GPT 5.2· Kagi LLM Benchmark73.3%
19Claude Sonnet 4thinking· Kagi LLM Benchmark73%
20GPT 5· Kagi LLM Benchmark72.7%
21Qwen3 Max· Kagi LLM Benchmark72.5%
22Claude Opus 4.6· Kagi LLM Benchmark72.4%
23O3 Pro· Kagi LLM Benchmark72.1%
24Claude Opus 4.5· Kagi LLM Benchmark70.7%
25Gemini 2.5 Pro· Kagi LLM Benchmark70.3%
26GPT-5 Codex· Kagi LLM Benchmark70.3%
27GPT-5 mini· Kagi LLM Benchmark70.3%
28DeepSeek R1· Kagi LLM Benchmark69.4%
29Qwen 3 235B A22B· Kagi LLM Benchmark69.4%
30o4 Mini· Kagi LLM Benchmark67.6%
31OpenAI o3· Kagi LLM Benchmark67.6%
32Gemini 3.1 Flash Lite· Kagi LLM Benchmark67.2%
33GPT-6.2 Sol· Kagi LLM Benchmark67%
34Qwen3 Next 80B A3Bthinking· Kagi LLM Benchmark66.7%
35Ring 1t· Kagi LLM Benchmark66.4%
36Grok 4 Fastthinking· Kagi LLM Benchmark66.1%
37Maestro Reasoning· Kagi LLM Benchmark64.9%
38Kimi K2 Thinking· Kagi LLM Benchmark64.4%
39GPT 5.4· Kagi LLM Benchmark63.8%
40Kimi K2.5· Kagi LLM Benchmark63.8%
41Gemma 4 31B IT· Kagi LLM Benchmark63.5%
42Qwen Plus· Kagi LLM Benchmark63.3%
43GLM 5.2· Kagi LLM Benchmark62.6%
44GPT-5 nano· Kagi LLM Benchmark62.2%
45Grok 3· Kagi LLM Benchmark61.3%
46Grok 3 Mini· Kagi LLM Benchmark61.3%
47Cogito V2 Preview Llama 405B· Kagi LLM Benchmark61.1%
48GLM 5.2thinking· Kagi LLM Benchmark60%
49GLM 4.5V· Kagi LLM Benchmark59.8%
50Claude Opus 4· Kagi LLM Benchmark59.6%

Showing 1–50 of 123 results.

Tokens per task, input and output together, as measured by the source. Comparable within this benchmark only.