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 · 4× 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	MiniMax M3 2026.07	427B	NVFP4	425K	9.10
2	Step-3.5-Flash 2026.03	199.4B	FP8	256K	9.10
3	DeepSeek-V4-Flash-0731 2026.08	304.2B	FP8	172K	8.98
4	Mistral-Medium-3.5-128B 2026.07	127.7B	FP16	256K	8.90
5	MiniMax-M2 2025.12	228.7B	FP8	192K	8.83
6	Qwen3.5-397B-A17B 2026.02	397B	NVFP4	256K	8.81
7	Qwen3-Coder-Next 2026.03	80B	FP16	256K	8.70
8	Qwen3.5-122B-A10B 2026.02	122B	FP16	256K	8.47
9	gpt-oss-120b 2025.08	116.8B	FP16	128K	8.30
10	Nemotron 3 Ultra 2026.06	550B	NVFP4	500K	8.00
11	MiniMax-M2.7 2026.04	228.7B	FP8	200K	7.96
12	Mistral-Small-4-119B-2603 2026.07	119.4B	FP16	460K	7.80
13	Step-3.7-Flash 2026.06	201.4B	FP8	256K	7.56
14	Mistral Medium 3.5 2026.04	128B	FP16	256K	7.50
15	MiniMax-M2.1 2026.02	228.7B	FP8	192K	7.40
16	DeepSeek-V4-Flash 2026.07	284B	FP8	382K	7.03
17	Leanstral 1.5 2026.06	119B	FP16	256K	7.00
18	Ling 2.6 Flash 2026.04	104B	FP16	256K	7.00
19	Nemotron 3 Super 2026.03	120B	FP16	976K	7.00
20	gpt-oss-safeguard-120b 2025.10	120.4B	FP16	128K	6.80

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.