List your company

Task success rate on DMV-Bench MC J=50 1.0

68.3Task Success RateDualMem
-2.73215.70934.1552.591Jun 25, 2026

Evaluation results

MethodLinks
DualMemBack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.0668.3
DualMemBack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.0665.1
M2ABack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.0664.7
CaptionBack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.0658.8
M2ABack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.0658.7
CaptionBack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.0647.7
WorldMMBack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.0637
MMABack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.0636.9
MMABack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.0635.6
WorldMMBack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.0630.9
NoMemoryBack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.060
TextOnlyBack-end=Qwen2.5-VL-7B, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4072026.060
NoMemoryBack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.060
TextOnlyBack-end=Gemini 2.5 Flash, Monte Carlo range (r)=1–49, Probe counts (nr)=2,4492026.060