CodeARC - Anonymized: leaderboard

Metric: Success Rate (%; anonymized function names, 1,114 functions). Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.

Top models

#ModelScore
1O3 Mini52.7
2O1 Mini47.7
3DeepSeek R141.3
4Claude 3.7 Sonnet33.8
5DeepSeek V329.5
6GPT-4o28.7
7Llama 3.1 405B Instruct26.7
8Qwen 2.5 72B Instruct21.6
9Llama 3.1 70B Instruct20
10Llama 3.2 90B Vision Instruct19.7
11GPT-4o Mini18.5
12Llama 3.2 11B Instruct16.1
13Qwen 2.5 7B Instruct15.8
14QwQ-32B15.4
15Mixtral 8x22B Instruct (v0.1)15

Interactive version: theaggregate.ai/benchmark?slug=codearc-anonymized · How It Works · Data refreshed daily, snapshot 2026-09-24.