GAIA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GAIA
47.17Success Rate (Level 1)
10
GAIA Plan Synthetic (test)
98Accuracy
10
GAIA Plan Human-annotated (test)
93Accuracy
10
GAIA Search 2
82.71Pass@1
9
GAIA
0.635Sim
9
GAIA Out-of-Domain (test)
31.1Accuracy
9
GAIA
46.3GAIA Level 1 Score
9
GAIA Text
59L1 Accuracy
9
GAIA
62.1Accuracy (Text)
8
GAIA Level 2
42.9Success Rate
8
GAIA Level 1
66.7Success Rate
8
GAIA Round 1
25Level 1 Evo. Pass@1
8
GAIA
49.2Performance Score
8
GAIA (test)
49.2Mean Correct Rate
8
GAIA
86.36Research And Data Score
8
GAIA
73.33Task Completion Rate
8
GAIA
58.33Task Completion Rate
8
GAIA
62.96Task Completion Rate
8
GAIA
65.91Task Completion Rate
8
GAIA (test)
63.11Pass@3
8
GAIA level2 Text-only
57.14Accuracy
8
GAIA Level 3
2Correct Samples
7
GAIA Level 2
22Number Correct
7
GAIA Level 1
24Correct Samples
7
GAIA 68 curated tasks
67.6Accuracy
7