Deploy Open Models on GPUs

Choose a GPU setup and a capability to find the best open models that fit, and deploy one as an API.

Agent · 1× NVIDIA RTX PRO 6000

Ranked by benchmark scores for Tool calling, multi-step planning and environment interaction

#	Model	Params	Precision	Max context	Score (0–10)
1	UI-Mate-democua-27B 2026.08	27.4B	FP16	256K	10.00
2	UI-Mate-27B 2026.08	27.4B	FP16	256K	10.00
3	UI-Mate-9B 2026.08	9.4B	FP16	256K	10.00
4	Qwen3.8-27B 2026.08	27.8B	FP16	256K	9.59
5	Qwen3.6-27B 2026.04	27.8B	FP16	256K	9.16
6	Mistral-Medium-3.5-128B 2026.07	127.7B	NVFP4	65K	8.90
7	Kimi-VL-A3B-Thinking-2506 2026.01	16.4B	FP16	128K	8.75
8	Fara1.5-27B 2026.07	27.4B	FP16	256K	8.70
9	Fara1.5-9B 2026.07	9.4B	FP16	256K	8.70
10	Fara1.5-4B 2026.07	4.5B	FP16	256K	8.70
11	Qwen-AgentWorld-35B-A3B 2026.06	34.7B	FP16	69K	8.70
12	Qwen3-Coder-Next 2026.03	80B	FP8	80K	8.70
13	Qwen3.5-35B-A3B 2026.02	35B	FP16	256K	8.47
14	Qwen3.5-27B 2026.02	27B	FP16	256K	8.47
15	Qwen3.5-122B-A10B 2026.02	122B	NVFP4	246K	8.47
16	gpt-oss-120b 2025.08	116.8B	NVFP4	90K	8.30
17	Fara-7B 2026.08	8.3B	FP16	125K	7.80
18	Mistral-Small-4-119B-2603 2026.07	119.4B	NVFP4	84K	7.80
19	Kimi-VL-A3B-Thinking 2026.01	16.4B	FP16	128K	7.80
20	Gemma 4 12B 2026.06	12B	FP16	256K	7.69

Deployment verified on HyperAI

Scores aggregate public benchmarks for the selected capability.
Fits usable VRAM with weights plus KV cache at 64K+ context, at the best precision.