Hardware

RTX 6000 Ada

RTX 6000 Ada has 48 GB GDDR6 and 960 GB/s. See which open models fit, at which quant, and roughly how fast. Figures below use the default memory (48 GB), 32 GB system RAM, and DDR5-5600 for anything that spills. Updated 2026-10-24.

Memory48 GB GDDR6

Bandwidth960 GB/s

Kindgpu

Year2023

48GB GDDR6, 960 GB/s. Source.

RSS for new models on this device

Models that fit

Model	Verdict	Quant	Context	tok/s
DeepSeek R1 Distill Llama 70B 70.6B	Tight fit	UD-Q4_K_XL	8K	12–16
Llama 3.3 70B Instruct 70.6B	Tight fit	UD-Q4_K_XL	8K	12–16
Hermes 4 70B 70.6B	Tight fit	UD-Q4_K_XL	8K	12–16
Qwen2.5 72B Instruct 72.7B	Tight fit	IQ4_XS	8K	12–17
Qwen3 Next 80B A3B Thinking 79.7B	Tight fit	UD-Q4_K_XL	4K	248–338
Qwen3 Next 80B A3B Instruct 79.7B	Tight fit	UD-Q4_K_XL	4K	247–337
Qwen3 Coder Next 79.7B	Tight fit	UD-Q4_K_S	4K	64–88
Mistral Small 4 119B 2603 119B	Tight fit	UD-IQ3_S	8K	64–87
Mistral Medium 3.5 128B 125B	Runs	UD-IQ2_XXS	8K	14–19
Llama 4 Scout 17B 16E Instruct 108B	Tight fit	UD-IQ3_XXS	4K	50–68
Qwen3.5 122B A10B MTP 125B	Tight fit	UD-Q2_K_XL	8K	145–198
Qwen2.5 32B Instruct 32.8B	Runs	Q8_0	8K	14–19
Qwen2.5 Coder 32B Instruct 32.8B	Runs	Q8_0	8K	14–19
DeepSeek R1 Distill Qwen 32B 32.8B	Runs	Q8_0	8K	14–19
Qwen3 32B 32.8B	Tight fit	UD-Q8_K_XL	8K	13–17
Qwen3 VL 32B Instruct 32.8B	Tight fit	UD-Q8_K_XL	8K	13–17
GLM 4 32B 0414 32.6B	Runs	UD-Q8_K_XL	8K	13–18
gemma 4 31B it 30.7B	Runs	UD-Q8_K_XL	8K	12–17
GLM 4.6V 107B	Tight fit	UD-Q2_K_XL	4K	50–69
Qwen3.8 27B 27.3B	Runs	UD-Q8_K_L	8K	18–25
Swift Qwen3.8 27B 27.3B	Runs	Q8_0	8K	18–24
gemma 3 27b it 27.0B	Runs	UD-Q8_K_XL	8K	15–20
medgemma 27b text it 27.0B	Runs	UD-Q8_K_XL	8K	15–20
Qwen3.6 27B 26.9B	Runs	UD-Q8_K_XL	8K	15–20
Qwen3.5 27B 26.9B	Runs	UD-Q8_K_XL	8K	15–20
Kwaipilot KAT Coder V2.5 Dev 34.7B	Runs	Q8_0	8K	77–105
Qwen3.8 35B A3B Distill 35.5B	Runs	Q8_0	8K	157–214
Cyber Tiel Coder 35B A3B GGUF MTP 35.5B	Runs	Q8_0 (est.)	8K	157–215
GLM 4.5 Air 110B	Tight fit	UD-IQ2_XXS	8K	50–68
Qwen3.5 35B A3B 34.7B	Runs	Q8_0	8K	157–214
Qwen3.6 35B A3B 34.7B	Runs	UD-Q8_K_XL	8K	151–206
Tiel Coder 35B A3B 34.7B	Runs	UD-Q8_K_XL	8K	151–206
NVIDIA Nemotron 3.5 Lightning 30B A3B 32.9B	Runs	UD-Q8_K_XL	8K	133–182
Devstral Small 2 24B Instruct 2512 23.6B	Runs	UD-Q8_K_XL	8K	17–24
Mistral Small 3.2 24B Instruct 2506 23.6B	Runs	UD-Q8_K_XL	8K	17–24
Magistral Small 2509 23.6B	Runs	UD-Q8_K_XL	8K	17–24
NVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning 31.6B	Runs	UD-Q8_K_XL	8K	139–190
Nemotron 3 Nano 30B A3B 31.6B	Tight fit	UD-Q8_K_XL	8K	123–168
GLM 4.7 Flash 29.9B	Runs	UD-Q8_K_XL	8K	66–90
Qwen3 Coder 30B A3B Instruct 30.5B	Runs	UD-Q8_K_XL	8K	122–166

105 more models in the catalog do not fit this default profile, or ranked below the first 40 that do.

Example: Tight fit fully on your RTX 6000 Ada at UD-Q4_K_XL, 8K context, about 12–16 tok/s. Weights 39.7 GB plus 2.50 GB of KV cache — 42.9 GB of 45.1 GB usable. Weights shown in the verdict use GiB (the same 1024³ counting as GPU memory).