CRMArena - KQA (ReAct): leaderboard

Metric: F1 (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.

Top models

#ModelScore
1DeepSeek R161.2
2O1 (2024-12-17)58.8
3Llama 3.1 405B Instruct55.3
4GPT-4o Mini52.8
5GPT-4o (2024-08-06)50.2
6Llama 3.1 70B Instruct48.7
7Claude 3.5 Sonnet (20240620)45
8Claude 3 Sonnet (20240229)16
9Llama 3.1 8B Instruct4

Interactive version: theaggregate.ai/benchmark?slug=crmarena-kqa-react · How It Works · Data refreshed daily, snapshot 2026-10-09.