List your company

Theory of Mind on FANToM

95AccuracyDITTO
-3.59222.00447.673.196May 19, 2026

Evaluation results

MethodLinks
DITTOBackbone=Qwen3-VL-8B-Instruct2026.0595
GRPOBackbone=Qwen3-VL-8B-Instruct2026.0594
GPT-5.42026.0590
HumanLM-8B2026.0578
Qwen3-VL-8B-InstructRole=Base2026.0578
OSCToM-8BParams=8B, Inference Protocol=direct model inference2026.0576
GPT-5-nano2026.0572
Llama-3.1-8B-BaseParams=8B, Inference Protocol=direct model inference2026.0566
Qwen2.5-14BParams=14B, Inference Protocol=direct model inference2026.0554.5
Phi-3-Medium-14BParams=14B, Inference Protocol=direct model inference2026.0551
Qwen2.5-32BParams=32B, Inference Protocol=direct model inference2026.0546.5
Gemma-2-27BParams=27B, Inference Protocol=direct model inference2026.0538.5
Mistral-NeMo-12BParams=12B, Inference Protocol=direct model inference2026.0537.5
ExploreToMParams=8B, Inference Protocol=iterative A* heuristic search2026.050.2