Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Chatbot Evaluation research benchmarks · SOTA2 Research

Chatbot Evaluation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

AI Barometer Estonian Chatbot Arena 19.02.2026

DeepSeek-V3 (0324)

1,457Score

20

Mar 4, 2026

MT-Bench High-Disagreement (Top 20%)

8.72Human Score

13

Mar 10, 2026

MT-Bench Overall

8.17Human Score

13

Mar 10, 2026

Vicuna benchmark

13,481Elo Rating

8

Feb 26, 2026

71.64Overall Score

6

May 19, 2026

Qwen3-32B + RLBFF training

9.5Score (GPT-4-Turbo)

6

May 19, 2026

57.4ArenaHard v2 Score

6

May 19, 2026

13.88Win Rate

3

May 22, 2026

Popular datasets

MT-Bench ArenaHard Vicuna WildBench AI Barometer

Follow for updates

Join SOTA2 Contact