What AI models can a RTX 5880 Ada Generation run?

2031 of 2118 indexed open-weight models fit a RTX 5880 Ada Generation at 16,384 context with f16 KV cache, the largest being HuatuoGPT-o1-72B at IQ4_NL. That covers text, vision-language, image, video and speech models.

How much usable memory does a RTX 5880 Ada Generation actually have?

Its nameplate is 48 GB, but about 44.64 GiB is available to a model once driver and compositor overhead is accounted for.

Is a RTX 5880 Ada Generation fast for local AI?

Its memory bandwidth is 960 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 5880 Ada Generation

RTX 5880 Ada Generation has 48 GB of VRAM at 960 GB/s — about 44.64 GiB usable after driver and compositor overhead. 2031 of 2118 indexed models fit at 16K context with f16 KV.

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

Memory

48 GB

GDDR6

Bandwidth

960 GB/s

384-bit bus

Tensor FP16

277 TF

dense

TDP

285 W

$6999 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

text 1743vision language 184image 2video 16audio tts 21embedding 26audio asr 39

What fits at 16K context

largest quantization that fits, per model · 2031 of 2118 indexed

Model	Best quant	Params	Weights●	KV●	Total in memory◐	Headroom◐	tok/s◐
HuatuoGPT-o1-72B	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Rombo-LLM-V3.0-Qwen-72b	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
MiroThinker-v1.0-72B	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Qwen2.5-72B	I1-IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Kimi-Dev-72B	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
KAT-Dev-72B-Exp	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Homer-v1.0-Qwen2.5-72B	IQ4_NL	72.7B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Qwen2.5-VL-72B-Instruct	IQ4_NL	73.4B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
UI-TARS-72B-DPO	IQ4_NL	73.4B	38.48 GiB	5.00 GiB	44.61 GiB	0.03 GiB	13±22%
Qwen3.5-122B-A10BMoE	Q2_K	125B	43.21 GiB	0.38 GiB	44.61 GiB	0.03 GiB	71±37%
GLM-4.5VMoE	I1-Q2_K	108B	40.61 GiB	2.88 GiB	44.51 GiB	0.13 GiB	40±37%
Phi-3.5-MoE-instructMoEKV unresolved	Q8_0	41.9B	41.44 GiB	2.00 GiB	44.45 GiB	0.19 GiB	33±37%
Qwen3.6-35B-A3B-REAM-160-ru-agentMoE	BF16	23.6B	43.09 GiB	0.31 GiB	44.41 GiB	0.23 GiB	62±37%
command-r-35b-writer-v2	I1-Q5_K_M	35.0B	23.29 GiB	20.00 GiB	44.40 GiB	0.24 GiB	13±22%
Assistant_Pepe_70B	Q4_K_S	70.6B	38.22 GiB	5.00 GiB	44.35 GiB	0.29 GiB	13±22%
Mistral-Medium-3.5-128B	IQ2_XXS	128B	37.65 GiB	5.50 GiB	44.31 GiB	0.33 GiB	13±22%
NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16MoE	Q4_K_S	75.4B	43.15 GiB	0.00 GiB	44.22 GiB	0.42 GiB	111±37%
HarmonicHarlequin_v5-20B	I1-IQ2_M	33.3B	10.67 GiB	32.50 GiB	44.21 GiB	0.43 GiB	13±22%
GLM-4.6VMoE	IQ2_M	108B	40.26 GiB	2.88 GiB	44.16 GiB	0.48 GiB	40±37%
Qwen3-53B-A3B-2507-THINKING-TOTAL-RECALL-v2-MASTER-CODERMoE	I1-Q6_K	53.0B	40.54 GiB	2.63 GiB	44.15 GiB	0.49 GiB	38±37%
Huihui-Qwen3-Coder-Next-abliteratedMoE	Q4_0	79.7B	42.78 GiB	0.38 GiB	44.15 GiB	0.49 GiB	81±37%
ALIA-40b-fc-2606	Q8_0	40.4B	40.02 GiB	3.00 GiB	44.13 GiB	0.51 GiB	13±22%
ALIA-40b-instruct-2606	Q8_0	40.4B	40.02 GiB	3.00 GiB	44.13 GiB	0.51 GiB	13±22%
OYM-Qimi-122B-A10B-K2.6MoE	I1-Q2_K	125B	42.67 GiB	0.38 GiB	44.07 GiB	0.57 GiB	72±37%
Qwopus3.5-122B-A10B-Kimi-K2.6-destill-healed-abliteratedMoE	Q2_K	123B	42.66 GiB	0.38 GiB	44.07 GiB	0.57 GiB	72±37%
Llama-4-Scout-17B-16E-InstructMoEKV unresolved	Q2_K	109B	40.03 GiB	3.00 GiB	44.06 GiB	0.58 GiB	40±37%
CalmeRys-78B-Orpo-v0.1	I1-Q3_K_M	78.0B	37.54 GiB	5.38 GiB	44.05 GiB	0.59 GiB	13±22%
calme-2.3-rys-78b	Q3_K_M	78.0B	37.54 GiB	5.38 GiB	44.05 GiB	0.59 GiB	13±22%
GLM-4.5-Air-REAP-82B-A12BMoE	Q3_K_L	81.9B	40.10 GiB	2.88 GiB	44.00 GiB	0.64 GiB	37±37%
Mixtral_34Bx2_MoE_60BMoE	Q5_K_M	60.8B	39.03 GiB	3.75 GiB	43.86 GiB	0.78 GiB	7±37%
Apertus-70B-Instruct-2509	Q4_K_S	70.6B	37.67 GiB	5.00 GiB	43.85 GiB	0.79 GiB	13±22%
Qwen3.5-88BMoE	I1-Q3_K_L	87.7B	42.43 GiB	0.38 GiB	43.84 GiB	0.80 GiB	65±37%
Qwen3-Coder-Next-Opus-4.6-Reasoning-DistilledMoE	Q4_K_S	—	42.37 GiB	0.38 GiB	43.73 GiB	0.91 GiB	81±37%
WizardLM-Uncensored-SuperCOT-StoryTelling-30b	Q4_K_M	32.5B	18.27 GiB	24.38 GiB	43.72 GiB	0.92 GiB	13±22%
Wizard-Vicuna-30B-Uncensored	I1-Q4_K_M	32.5B	18.27 GiB	24.38 GiB	43.72 GiB	0.92 GiB	13±22%
archangel_sft-kto_llama30b	I1-Q4_K_M	32.5B	18.27 GiB	24.38 GiB	43.72 GiB	0.92 GiB	13±22%
Meta-Llama-3-70B-Instruct	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.71 GiB	0.93 GiB	13±22%
Maenad-70B	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
DeepSeek-R1-Distill-Llama-70B-Uncensored-v2-Unbiased-Reasoner	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
calme-2.4-llama3-70b	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
calme-2.2-llama3-70b	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Rombos-LLM-70b-Llama-3.3	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
L3.3-Electra-R1-70b	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
L3.3-70B-Magnum-v4-SE	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Latxa-Llama-3.1-70B-Instruct-v2	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Llama-3.3_70_b_uncensored_continued	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Llama-3.3-70B-Instruct-abliterated	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Strawberrylemonade-L3-70B-v1.2	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
grok-oss-Revenant-70B	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Llama-3.1-Nemotron-70B-Instruct-HF	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
L3.3-70B-Euryale-v2.3	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Hermes-4-70B-heretic	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Hermes-4-70B	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Llama-3.3-70B-Instruct	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Llama-3.1-70B	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Hermes-3-Llama-3.1-70B	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Anubis-70B-v1.2	Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
Golem-70B-v1b	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
DeepSeek-R1-Distill-Llama-70B-abliterated	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±22%
DeepSeek-R1-Distill-Llama-70B-heretic	I1-Q4_K_S	70.6B	37.58 GiB	5.00 GiB	43.70 GiB	0.94 GiB	13±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 5880 Ada Generation run?: 2031 of 2118 indexed open-weight models fit a RTX 5880 Ada Generation at 16,384 context with f16 KV cache, the largest being HuatuoGPT-o1-72B at IQ4_NL. That covers text, vision-language, image, video and speech models.
How much usable memory does a RTX 5880 Ada Generation actually have?: Its nameplate is 48 GB, but about 44.64 GiB is available to a model once driver and compositor overhead is accounted for.
Is a RTX 5880 Ada Generation fast for local AI?: Its memory bandwidth is 960 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.