Benchgen

DROP — Results

RankModelScore
1deepseek-v30.916
2claude-3-5-sonnet0.871
3gpt-4-turbo0.86
4llama-3-1-405b-instruct0.848
5gpt-4o0.834
6claude-3-5-haiku0.831
7claude-3-opus0.831
8gpt-4-06130.809
9gpt-4o-mini0.797
10llama-3-1-70b-instruct0.796
11phi-40.755
12gemini-1-5-pro0.749
13llama-3-1-8b-instruct0.595

DROP — Phases

1 phaseActive

Phase 1·

DROP

Aug 3, 2026
Current