Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Robotics Multimodal Understanding on Cosmos-ER
Loading...
74.9
Score
Cosmos-Reason2 32B
36.524
46.487
56.45
66.413
Jun 1, 2026
Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Score
Cosmos-Reason2 32B
Model Size=32B
2026.06
74.9
Cosmos 3 Super
Model Size=32B
2026.06
74.1
Cosmos-Reason2 8B
Model Size=8B
2026.06
71.2
Cosmos 3 Nano
Model Size=8B
2026.06
69.7
Qwen3-VL 32B
Model Size=32B
2026.06
61.3
Gemini 3.1 Pro
Access=Closed Model
2026.06
61.1
Cosmos-Reason2 2B
Model Size=2B
2026.06
59
Qwen3-VL 8B
Model Size=8B
2026.06
56.9
Cosmos 3 Edge
Model Size=2B
2026.06
56.6
MiMo-Embodied 7B
Model Size=7B
2026.06
55.6
RynnBrain 8B
Model Size=8B
2026.06
54.9
Gemma-4 31B
Model Size=31B
2026.06
54.3
Qwen3-VL 2B
Model Size=2B
2026.06
48.9
RynnBrain 2B
Model Size=2B
2026.06
48
Gemma-4 E4B
Model Size=E4B
2026.06
44.3
Gemma-4 E2B
Model Size=E2B
2026.06
38
Feedback
Search any
task
Search any
task