What AI models can a RTX 5000 Ada Generation run?

2022 of 2118 indexed open-weight models fit a RTX 5000 Ada Generation at 8,192 context with q8_0 KV cache, the largest being Bernini-R at Q8_0. That covers text, vision-language, image, video and speech models.

How much usable memory does a RTX 5000 Ada Generation actually have?

Its nameplate is 32 GB, but about 29.76 GiB is available to a model once driver and compositor overhead is accounted for.

Is a RTX 5000 Ada Generation fast for local AI?

Its memory bandwidth is 576 GB/s, and that figure — not teraflops — is what governs token generation speed. Capacity decides what you can run; bandwidth decides how fast it runs.

NVIDIA · workstation

RTX 5000 Ada Generation

RTX 5000 Ada Generation has 32 GB of VRAM at 576 GB/s — about 29.76 GiB usable after driver and compositor overhead. 2022 of 2118 indexed models fit at 8K context with q8_0 KV.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed

Memory

32 GB

GDDR6

Bandwidth

576 GB/s

256-bit bus

Tensor FP16

261 TF

dense

TDP

250 W

$4000 MSRP

Context4K 8K 16K 32K 64K 128K

KV cachef16 q8_0 q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context

video 16text 1737vision language 181image 2embedding 26audio tts 21audio asr 39

What fits at 8K context

largest quantization that fits, per model · 2022 of 2118 indexed

Model	Best quant	Params	Weights●	KV●	Total in memory◐	Headroom◐	tok/s◐
Bernini-R	Q8_0	14.3B	28.71 GiB	0.00 GiB	29.76 GiB	0.00 GiB	12±22%
Maenad-70B	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
DeepSeek-R1-Distill-Llama-70B-Uncensored-v2-Unbiased-Reasoner	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Rombos-LLM-70b-Llama-3.3	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
L3.3-Electra-R1-70b	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Latxa-Llama-3.1-70B-Instruct-v2	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Llama-3.3_70_b_uncensored_continued	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Llama-3.3-70B-Instruct-abliterated	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
grok-oss-Revenant-70B	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Llama-3.1-Nemotron-70B-Instruct-HF	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
L3.3-70B-Euryale-v2.3	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Hermes-3-Llama-3.1-70B	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Hermes-4-70B-heretic	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Llama-3.3-70B-Instruct	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Llama-3.1-70B	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Anubis-70B-v1.2	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Hermes-4-70B	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Golem-70B-v1b	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
DeepSeek-R1-Distill-Llama-70B-abliterated	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
DeepSeek-R1-Distill-Llama-70B-heretic	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
llama-3-firefunction-v2	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Legion-V2.1-LLaMa-70B	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Assistant_Pepe_70B	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Tess-R1-Limerick-Llama-3.1-70B	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Infinity-Instruct-7M-Gen-Llama3_1-70B	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
New-Dawn-Llama-3-70B-32K-v1.0	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Meta-Llama-3-70B-Instruct-abliterated-v3.5	I1-IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Athene-70B	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
L3.3-70B-Magnum-Diamond	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Meta-Llama-3-70B-Instruct	IQ3_XS	70.6B	27.29 GiB	1.33 GiB	29.75 GiB	0.01 GiB	12±22%
Salience-1.5-ProMoE	Q6_K_L	36.0B	28.66 GiB	0.08 GiB	29.75 GiB	0.01 GiB	69±37%
Qwable-v1MoE	Q6_K_L	36.0B	28.66 GiB	0.08 GiB	29.75 GiB	0.01 GiB	69±37%
T-SearchMoE	Q6_K_L	36.0B	28.66 GiB	0.08 GiB	29.75 GiB	0.01 GiB	69±37%
Melody1437-27B	Q3_K_M	27.8B	28.40 GiB	0.27 GiB	29.73 GiB	0.03 GiB	12±22%
Nous-Hermes-Llama2-70b	Q2_K	69.0B	27.27 GiB	1.33 GiB	29.72 GiB	0.04 GiB	12±22%
llama2_70b_chat_uncensored	Q2_K	69.0B	27.27 GiB	1.33 GiB	29.72 GiB	0.04 GiB	12±22%
Xwin-LM-70b-V0.1	Q2_K	69.0B	27.27 GiB	1.33 GiB	29.72 GiB	0.04 GiB	12±22%
Llama-2-70b-chat-hf	Q2_K	69.0B	27.27 GiB	1.33 GiB	29.72 GiB	0.04 GiB	12±22%
command-r-35b-writer-v2	I1-Q5_K_M	35.0B	23.29 GiB	5.31 GiB	29.71 GiB	0.05 GiB	12±22%
Qwen3-53B-A3B-2507-THINKING-TOTAL-RECALL-v2-MASTER-CODERMoE	I1-Q4_0	53.0B	28.02 GiB	0.70 GiB	29.71 GiB	0.05 GiB	42±37%
Qwen2.5-7B-Instruct-1M	F32	7.6B	28.38 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
DeepSeek-R1-Distill-Qwen-7B	F32	7.6B	28.38 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
UI-TARS-7B-DPO	F32	8.3B	28.38 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
Qwen2-7B-Instruct	F32	7.6B	28.38 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
Hercules-5.0-Qwen2-7B	F32	7.6B	28.38 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
Kepler-8B-Instruct-v2	F16	7.6B	28.37 GiB	0.23 GiB	29.66 GiB	0.10 GiB	12±22%
MiniCPM-o-2_6	F32	8.7B	28.37 GiB	0.23 GiB	29.65 GiB	0.11 GiB	12±22%
Qwen3-42B-A3B-2507-Thinking-Abliterated-uncensored-TOTAL-RECALL-v2-Medium-MASTER-CODERMoE	I1-Q5_K_M	42.4B	28.05 GiB	0.56 GiB	29.60 GiB	0.16 GiB	47±37%
Mixtral-8x22B-Instruct-v0.1MoE	IQ1_S	141B	27.61 GiB	0.93 GiB	29.60 GiB	0.16 GiB	21±37%
Mixtral-8x22B-v0.1MoE	IQ1_S	141B	27.61 GiB	0.93 GiB	29.60 GiB	0.16 GiB	21±37%
CalmeRys-78B-Orpo-v0.1	I1-IQ2_XS	78.0B	27.00 GiB	1.43 GiB	29.55 GiB	0.21 GiB	12±22%
calme-2.3-rys-78b	IQ2_XS	78.0B	27.00 GiB	1.43 GiB	29.55 GiB	0.21 GiB	12±22%
Huihui-GLM-4.7-Flash-abliterated-57BMoE	IQ4_XS	57.3B	27.93 GiB	0.56 GiB	29.53 GiB	0.23 GiB	48±37%
Qwen3.5-88BMoE	I1-Q2_K_S	87.7B	28.30 GiB	0.10 GiB	29.42 GiB	0.34 GiB	61±37%
Kimi-Linear-48B-A3B-InstructMoE	Q4_K_L	49.1B	28.26 GiB	0.13 GiB	29.39 GiB	0.37 GiB	12±22%
DeepCoder-14B-Preview	BF16	14.8B	27.52 GiB	0.80 GiB	29.36 GiB	0.40 GiB	12±22%
SuperNova-Medius	F16	14.8B	27.52 GiB	0.80 GiB	29.36 GiB	0.40 GiB	12±22%
Qwen2.5-14B-Instruct-1M	F16	14.8B	27.52 GiB	0.80 GiB	29.36 GiB	0.40 GiB	12±22%
OpenCodeReasoning-Nemotron-14B	BF16	14.8B	27.52 GiB	0.80 GiB	29.36 GiB	0.40 GiB	12±22%
Qwen2.5-Coder-14B-Instruct-abliterated	F16	14.8B	27.52 GiB	0.80 GiB	29.36 GiB	0.40 GiB	12±22%

●From the file◐Predictedwhat these mean

Speed is modeled, not measured: decode is memory-bandwidth bound, so tokens per second is bytes read per token against achievable bandwidth. Mixture-of-experts models carry a wider band because only the routed experts are read each step, and few have been measured publicly.

Questions people ask

What AI models can a RTX 5000 Ada Generation run?: 2022 of 2118 indexed open-weight models fit a RTX 5000 Ada Generation at 8,192 context with q8_0 KV cache, the largest being Bernini-R at Q8_0. That covers text, vision-language, image, video and speech models.
How much usable memory does a RTX 5000 Ada Generation actually have?: Its nameplate is 32 GB, but about 29.76 GiB is available to a model once driver and compositor overhead is accounted for.
Is a RTX 5000 Ada Generation fast for local AI?: Its memory bandwidth is 576 GB/s, and that figure — not teraflops — is what governs token generation speed. Capacity decides what you can run; bandwidth decides how fast it runs.