LOCAL AIregistry
‹ all GPUs

RTX PRO 4000 Blackwell

NVIDIA · 24 GB · 672 GB/s
Top 2 recipes
Recommended
Qwen3.8-27B
EXL3 4 bpwTabbyAPI256K context

Best for Coding agents, long documents, everyday assistant work.

72 tok/sdecode
509 tok/sprefill
256Kcontext
✓ Thinks✓ Tools✓ Long context✓ Vision
Tested on this card · Sep 25, 2026
Fastest
Gemma 4 26B A4B
EXL3 5.1 bpwTabbyAPI64K context

Best for Fast multimodal work on 16–24 GB cards.

76 tok/sdecode
2,306 tok/sprefill
64Kcontext
✓ Thinks✓ Tools✓ Long context✓ Vision
Tested on this card · Sep 26, 2026