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

Multimodal Conversation research benchmarks · SOTA2 Research

Multimodal Conversation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

LLaVA-Bench Wild

102Score

78

May 27, 2026

GPT-4V (2023.11.06)

93.1LLaVA Bench Score

46

May 19, 2026

PCogAlignBench (LS2)

GRPO (Latent Action)

0.852LLM Judge Score

20

Feb 26, 2026

PCogAlignBench LS1

DAPO (Latent Action)

0.903LLM Judge Score

20

Feb 26, 2026

Dr.GRPO (Latent Action)

91.6LLM-as-a-Judge Score

20

Feb 26, 2026

Dr.GRPO (Latent Action)

95.3LLM-as-a-Judge Score

20

Feb 26, 2026

Multimodality Chatbot Arena

LLaMA-Adapter v2

1,023Elo Rating

8

Feb 26, 2026

LLaVA-Bench In-the-Wild v1 (test)

0.726General Score

7

Feb 26, 2026

Popular datasets

LLaVA-Bench PCogAlignBench MMRole Multimodality Chatbot Arena

Follow for updates

Join SOTA2 Contact