Open LLM Leaderboard v1 - ARC Challenge: leaderboard

Metric: Normalized accuracy (%) (25-shot). Source: huggingface.co. Saturation forecast: Estimated already saturated. 7252 models tracked.

Top models

#ModelScore
1free-evo-qwen72B-v0.8-re79.86
2Rhea-72B-v0.579.78
3Le_Triomphant-ECE-TW378.5
4luxia-21.4B-alignment-v1.277.73
5UNA-ThePitbull-21.4B-v277.73
6final_model_test_v277.73
7Smaug-72B-v0.176.02
8Myrrh_solar_10.7b_3.075.43
9Truthful_DPO_TomGrc_FusionNet_7Bx2_MoE_13B74.91
10LLaMAntino-3-ANITA-8B-Inst-DPO-ITA74.57
11UNA-SimpleSmaug-34B-v1beta74.57
12Luminex-34B-v0.274.49
13LogoS-7Bx2-MoE-13B-v0.274.4
14DARE_TIES_13B74.32
15MoE_13B_DPO74.32

Interactive version: theaggregate.ai/benchmark?slug=open-llm-leaderboard-v1-arc-challenge · How It Works · Data refreshed daily, snapshot 2026-09-23.