Software Engineering Arena - Model Arena: leaderboard

Software Engineering Arena model-only preference leaderboard for software-engineering task performance.

Metric: Elo Rating. Source: huggingface.co. 23 models tracked.

Top models

#ModelScoreOverall rank
1Llama 3.1 70B Instruct1002#561
2O3 Mini (High)1002#272 (O3 Mini)
3MiMo-V2-Flash1002#372
4DeepSeek V3.1 Terminus1002#220
5Grok Code Fast 11002#337
6Qwen Turbo1002#699
7Qwen Plus1002#385
8Gemini 2.5 Pro998#152
9GPT-4.1998#248
10Claude Haiku 4.5998#280
11Qwen 3 8B998#674
12Gemini 2.5 Flash Lite998#422
13Mixtral 8x22B Instruct998#728
14GLM-4.5998#275
15NVIDIA Nemotron Nano 9B V2998#823

Interactive version: theaggregate.ai/benchmark?slug=software-engineering-arena-model-arena · How It Works · Data refreshed daily, snapshot 2026-10-20.