Skip to content

Benchmark

GAIA

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

Models measured
71
Scores recorded
291
Sources
3
IQ weight
1.01

Score distribution

37%
Mean
32.2%
Median
28.6%
Std dev
2.3%
10th percentile
80.6%
90th percentile
0%
Min
0%94.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.01 for chat models, measured over 71 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
251Qwen3 32BLevel 3· GAIA Leaderboard data4.1%
252Qwen3 Next 80B A3B InstructLevel 3· GAIA Leaderboard data4.1%
253qwen-coderLevel 3· GAIA Leaderboard data3.8%
254QwQ 32BLevel 3· GAIA Leaderboard data3.8%
255DeepSeek R1 Distill Qwen 7BLevel 2· GAIA Leaderboard data3.8%
256Gemini 2.0 FlashLevel 2· GAIA Leaderboard data3.8%
257GPT-3.5Level 1· GAIA Leaderboard data3.8%
258DeepSeek R1 Distill Qwen 7B· GAIA Leaderboard data3.7%
259GPT 5· GAIA Leaderboard data3.7%
260GPT 5Level 2· GAIA Leaderboard data2.5%
261GPT-3.5· GAIA Leaderboard data2.4%
262GPT 4Level 2· GAIA Leaderboard data2.3%
263GPT-3.5Level 2· GAIA Leaderboard data2.3%
264Gemini 3.6 FlashLevel 1· GAIA Leaderboard data2.2%
265claudeLevel 3· GAIA Leaderboard data2%
266DeepSeek V3.1Level 3· GAIA Leaderboard data2%
267Qwen 2.5 Coder 32B InstructLevel 3· GAIA Leaderboard data2%
268Qwen 3.8 27BLevel 3· GAIA Leaderboard data2%
269Gemini 3.6 FlashLevel 2· GAIA Leaderboard data1.9%
270Gemini 3.6 Flash· GAIA Leaderboard data1.7%
271llama3.1Level 1· GAIA Leaderboard data1.1%
272llama3.1· GAIA Leaderboard data0.664%
273llama3.1Level 2· GAIA Leaderboard data0.629%
274Qwen 2 (1.5B)Level 2· GAIA Leaderboard data0.629%
275Qwen 2 (1.5B)· GAIA Leaderboard data0.332%
276Qwen 2.5 Coder 32B Instruct· GAIA Leaderboard data0.332%
277Qwen 3.8 27B· GAIA Leaderboard data0.332%
278claudeLevel 1· GAIA Leaderboard data0%
279Claude Haiku 4.5Level 3· GAIA Leaderboard data0%
280DeepSeek R1 Distill Qwen 7BLevel 3· GAIA Leaderboard data0%
281Gemini 2.0 FlashLevel 3· GAIA Leaderboard data0%
282Gemini 3.6 FlashLevel 3· GAIA Leaderboard data0%
283GPT 4Level 3· GAIA Leaderboard data0%
284GPT-3.5Level 3· GAIA Leaderboard data0%
285llama3.1Level 3· GAIA Leaderboard data0%
286Qwen 2 (1.5B)Level 1· GAIA Leaderboard data0%
287Qwen 2 (1.5B)Level 3· GAIA Leaderboard data0%
288Qwen 2.5 Coder 32B InstructLevel 1· GAIA Leaderboard data0%
289Qwen 2.5 Coder 32B InstructLevel 2· GAIA Leaderboard data0%
290Qwen 3.8 27BLevel 1· GAIA Leaderboard data0%
291Qwen 3.8 27BLevel 2· GAIA Leaderboard data0%

Showing 251–291 of 291 results.

Data credits: Steel.dev Agent Leaderboards, MIT (licence) · 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.