Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Abstract Reasoning on ConceptARC
Loading...
97.29
Accuracy
Qwen3 + DIARC
-2.81
23.1775
49.165
75.1525
Jun 25, 2026
Accuracy
Updated 29d ago
Evaluation Results
Method
Method
Links
Accuracy
Qwen3 + DIARC
#Params=3.6B
2026.06
97.29
Qwen3 + SFT only
#Params=3.6B
2026.06
96.46
Human (pass@1)
#Params=-
2026.06
73
o4-mini
#Params=N/A
2026.06
70.8
Gemini 2.5 Pro
#Params=N/A
2026.06
66
Minitron + DIARC
#Params=6.9B
2026.06
62.71
Minitron + SFT only
#Params=6.9B
2026.06
60.42
Claude Sonnet 4
#Params=N/A
2026.06
60.2
Llama-3.2 + DIARC
#Params=2.6B
2026.06
59.17
Llama-3.2 + SFT only
#Params=2.6B
2026.06
58.13
Qwen-2.5
#Params=7B
2026.06
26
GPT-4
#Params=N/A
2026.06
24
o4-mini (visual)
#Params=N/A
2026.06
8.1
Minitron
#Params=8B
2026.06
5.83
Qwen3
#Params=4B
2026.06
1.25
Llama-3.2
#Params=3B
2026.06
1.04
Feedback
Search any
task
Search any
task