Benchgen

MT-Bench — Results

RankModelScore
1qwen2-5-72b-instruct9.35
2gpt-4-turbo9.32
3llama-3-3-nemotron-super-49b-v19.17
4deepseek-v2-59.02
5claude-3-opus9
6gpt-4-06138.99
7hermes-3-70b8.99
8claude-3-5-sonnet8.97
9qwen2-5-7b-instruct8.75
10llama-3-1-70b-instruct8.7
11mistral-large-28.63
12qwen2-7b-instruct8.41
13mistral-small-3-24b-instruct8.35
14gemini-1-5-pro8.3
15ministral-8b-instruct8.3
16llama-3-1-nemotron-nano-8b-v18.1
17claude-3-sonnet8.03
18claude-3-haiku8
19llama-3-1-nemotron-70b-instruct0.9
M

MT-Bench — Leaderboard

1 phaseActive

MT-Bench Leaderboard

No self-reported entries.