Benchgen

MBPP — Results

RankModelScore
1twil-8b93.7
2llama-3-3-nemotron-super-49b-v10.913
3qwen2-5-coder-32b-instruct0.902
4qwen2-5-72b-instruct0.882
5qwen2-5-32b-instruct0.84
6qwen2-5-vl-32b-instruct0.84
7qwen2-5-coder-7b-instruct0.835
8qwen2-5-14b-instruct0.82
9qwen3-235b-a22b0.814
10phi-3-5-moe-instruct0.808
11qwen2-72b-instruct0.802
12qwen2-5-7b-instruct0.792
13llama-4-maverick0.776
14mistral-small-3-1-24b-instruct0.747
15qwen2-5-omni-7b0.732
16mistral-small-3-24b-base0.696
17phi-3-5-mini-instruct0.696
18llama-4-scout0.678
19qwen2-7b-instruct0.672

MBPP — Phases

1 phaseActive

Phase 1·

MBPP

Aug 3, 2026
Current