Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-agent Debate on MMLU-Pro Debate
Loading...
65.7
Success Rate (SR)
Concise
51.452
55.151
58.85
62.549
Apr 7, 2026
Success Rate (SR)
Cost
Updated 3mo ago
Evaluation Results
Method
Method
Links
Success Rate (SR)
Cost
Concise
Listener=Llama-70B
2026.04
65.7
847
HANDRAISER
Listener=Llama-70B
2026.04
65.7
806
Prompting
Listener=Llama-70B
2026.04
65.3
1,707
Generic
Listener=Llama-70B
2026.04
64.7
1,617
Random
Listener=Llama-70B
2026.04
62.3
1,617
HANDRAISER
Listener=Llama-8B
2026.04
58.3
782
Random
Listener=Llama-8B
2026.04
53.7
1,531
Prompting
Listener=Llama-8B
2026.04
53.7
1,690
Concise
Listener=Llama-8B
2026.04
52.3
829
Generic
Listener=Llama-8B
2026.04
52
1,531
Feedback
Search any
task
Search any
task