Best open-source models for image understanding
Compare the performance of the best open-source models for image understanding, side by side.

Pixtral-12B
48GB VRAM
0.88$-1.03$/hour
No analysis performed yet

InternVL2.5-1B
16GB VRAM
0.28$/hour
No analysis performed yet

InternVL2.5-2B
16GB VRAM
0.28$/hour
No analysis performed yet

InternVL2.5-4B
16GB VRAM
0.28$/hour
No analysis performed yet

InternVL2.5-8B
24GB VRAM
0.43$-0.69$/hour
No analysis performed yet

Llama-3.2-11B-Vision-Instruct
48GB VRAM
0.88$-1.03$/hour
No analysis performed yet

DeepSeek-Janus-Pro-1B
16GB VRAM
0.28$/hour
No analysis performed yet

DeepSeek-Janus-Pro-7B
24GB VRAM
0.43$-0.69$/hour
No analysis performed yet