Benchgen

Winogrande — Results

RankModelScore
1ernie-5-5-flash-base87.8
2gpt-4-06130.875
3mimo-v2-5-pro0.856
4command-r-plus0.854
5qwen2-72b-instruct0.851
6llama-3-1-nemotron-70b-instruct0.845
7gemma-2-27b0.837
8hermes-3-70b0.832
9qwen2-5-32b-instruct0.82
10phi-3-5-moe-instruct0.813
11qwen2-5-coder-32b-instruct0.808
12gemma-2-9b0.806
13mistral-nemo-instruct0.768
14ministral-8b-instruct0.753
15granite-3-3-8b-base0.744
16qwen2-5-coder-7b-instruct0.729
17gemma-3n-e4b-litert-preview0.717
18gemma-3n-e4b0.717
19phi-3-5-mini-instruct0.685
20phi-4-mini0.67
21gemma-3n-e2b-litert-preview0.668
22gemma-3n-e2b0.668
23ernie-4-50.513

Winogrande — Leaderboard

1 phaseActive

Winogrande Leaderboard

No self-reported entries.