Desktop GPU · NVIDIA

What AI models can a GeForce RTX 5090 run?

32 GB puts the 70B class within comfortable reach and leaves room for the context window that makes those models worth running. At 1792 GB/s there is enough bandwidth to keep generation responsive for anything that fits.

The only consumer card with 32 GB, and 1,792 GB/s of bandwidth, nearly twice a 3090. It draws up to 575 W and needs a power supply, case and cable to match. For language models alone it is a luxury over a used 3090, and for image and video work it is the fastest single card you can buy.

Memory32 GB

Bandwidth1792 GB/s

Usable for a model31.2 GB

Runtime backendCUDA

ArchitectureBlackwell

Power575 W

The short answer

Assuming an 8k context window and default settings, these are the models worth downloading first.

Best all-rounder

Gemma 4 31B Instruct

Q6_K · 27.61 GB · about 48.5 tokens/s

Best for code

Qwen3 32B

Q6_K · 26.86 GB · about 49.9 tokens/s

Largest that still runs well

Mixtral 8x7B Instruct V0.1

46.7B parameters · Q4_K_M · about 93 tokens/s

Fastest useful answer

LFM2.5 8B A1B

about 464.2 tokens/s · 8.81 GB

See how fast it feels

GeForce RTX 5090 running Gemma 4 31B Instruct at Q6_K

Wait for the first word201 ms

Then writes at48.5 tok/s

Whole answer3.8 s

YouWhy does my model use more memory when the conversation gets longer?

Model

Simulated from our estimate at a 512-token question, not a recording. Assumes nothing else is competing for the GPU.

Where it sits

With 32 GB it belongs to the 32 GB tier. The 32 GB guide explains what that tier runs and where it runs out, and the bandwidth guide explains why 1792 GB/s decides how fast it answers.

Our indicative price for this card is $2,200 new, reviewed 2026-09-01. That is 1.5 GB of memory and 81 GB/s of bandwidth for every $100. Prices move, so treat these as a place to start a search and not a quote.

Devices in a similar price band, fastest memory first:

Device	Price	Memory	Bandwidth	Best all-round pick
GeForce RTX 4090	$1,900	24 GB	1008 GB/s	Qwen3.5 27B, 35.5 t/s

Can an RTX 5090 run it?

A direct answer for the models people ask about most, each with the settings we would use.

Qwen2.5 7B Instruct Llama 3.1 8B Instruct Qwen3 8B Qwen3.5 9B Gemma 4 12B Instruct Qwen3 14B Phi 4 GPT OSS 20B Qwen3.5 27B Gemma 4 31B Instruct Qwen3 32B Qwen3 30B A3B Llama 3.3 70B Instruct GPT OSS 120B

Every model, scored on this device

Each row uses the highest-quality quantization that both fits and stays conversational. Speed is a single-stream estimate at 8k context.

Model	Params	Verdict	Download	Memory used	Speed	Max context
Gemma 4 31B Instruct Gemma	31.27B	Runs great	Q6_K	27.61 GB	48.5 t/s	16k
Qwen3.5 27B Qwen	27.78B	Runs great	Q8_0	27.32 GB	49.1 t/s	64k
Qwen3 32B Qwen	32.76B	Runs great	Q6_K	26.86 GB	49.9 t/s	16k
Granite 4.1 30B Granite	28.87B	Just fits	Q8_0	30.97 GB	43.1 t/s	8k
Qwen2.5 32B Instruct Qwen	32.76B	Runs great	Q6_K	26.86 GB	49.9 t/s	16k
Qwen2.5 Coder 32B Instruct Qwen	32.76B	Runs great	Q6_K	26.86 GB	49.9 t/s	16k
Gemma 3 27B Instruct Gemma	27.43B	Runs great	Q8_0	28.68 GB	46.7 t/s	32k
Gemma 4 26B A4B Instruct Gemma	25.81B (3.74B active)	Runs great	Q8_0	26.72 GB	169.1 t/s	64k
Qwen3.5 35B A3B Qwen	35.95B (2.9B active)	Runs great	Q6_K	27.62 GB	318.7 t/s	128k
GLM 4.7 Flash GLM	31.22B (3.66B active)	Runs great	Q6_K	24.35 GB	240.5 t/s	128k
Qwen3 30B A3B Qwen	30.53B (3.34B active)	Just fits	Q8_0	30.99 GB	190.2 t/s	8k
Mistral Small 24B Instruct 2501 Mistral	23.57B	Runs great	Q8_0	24.35 GB	55.1 t/s	32k
Gemma 4 12B Instruct Gemma	11.96B	Runs great	Q8_0	13.55 GB	100.3 t/s	128k
Qwen3 14B Qwen	14.77B	Runs great	Q8_0	15.54 GB	87.3 t/s	32k
GPT OSS 20B GPT-OSS	20.91B (4.18B active)	Runs great	MXFP4	11.55 GB	320.2 t/s	128k
Qwen3 Coder 30B A3B Instruct Qwen	30.53B (3.34B active)	Just fits	Q8_0	30.99 GB	190.2 t/s	8k
Qwen2.5 14B Instruct Qwen	14.77B	Runs great	Q8_0	15.79 GB	85.9 t/s	32k
Qwen2.5 Coder 14B Instruct Qwen	14.77B	Runs great	Q8_0	15.79 GB	85.9 t/s	32k
Codestral 22B V0.1 Mistral	22.25B	Runs great	Q8_0	24.04 GB	55.9 t/s	32k
Qwen3.5 9B Qwen	9.65B	Runs great	Q8_0	9.25 GB	149.1 t/s	256k
Phi 3.5 MoE Instruct Phi	41.87B (6.64B active)	Runs great	Q5_K_M	29.05 GB	142 t/s	16k
Mixtral 8x7B Instruct V0.1 Mistral	46.7B (12.88B active)	Runs great	Q4_K_M	27.59 GB	93 t/s	32k
Phi 4 Phi	14.66B	Runs great	Q8_0	16 GB	84.7 t/s	16k
Llama 3.3 70B Instruct Llama	70.55B	Fits only at 2 to 3 bits	IQ3_XXS !	27.8 GB	48.4 t/s	16k
Qwen2.5 VL 72B Instruct Qwen	73.41B	Fits only at 2 to 3 bits	IQ3_XXS !	28.75 GB	46.7 t/s	8k
Gemma 3 12B Instruct Gemma	12.19B	Runs great	Q8_0	13.27 GB	102.4 t/s	128k
Qwen2.5 72B Instruct Qwen	72.71B	Fits only at 2 to 3 bits	IQ3_XXS !	28.5 GB	47.2 t/s	16k
Gemma 4 E4B Instruct Gemma	8B	Runs great	Q8_0	8.42 GB	163.6 t/s	128k
Granite 4.1 8B Granite	8.79B	Runs great	Q8_0	10.35 GB	132.5 t/s	128k
Qwen3 8B Qwen	8.19B	Runs great	Q8_0	9.02 GB	153.1 t/s	32k

Show the other 56 models

Mistral Nemo Instruct 2407	12.25B	Runs great	Q8_0	13.15 GB	103.8 t/s	64k
Qwen2.5 VL 7B Instruct	8.29B	Runs great	Q8_0	8.49 GB	162.7 t/s	64k
Qwen2.5 7B Instruct	7.62B	Runs great	Q8_0	7.83 GB	177.2 t/s	32k
Llama 3.1 8B Instruct	8.03B	Runs great	Q8_0	8.83 GB	156.5 t/s	128k
DeepSeek Coder V2 Lite Instruct	15.71B	Runs great	Q8_0	15.91 GB	262.6 t/s	128k
Gemma 2 9B Instruct	9.24B	Runs great	Q8_0	11.5 GB	118.7 t/s	8k
Qwen2.5 Coder 7B Instruct	7.62B	Runs great	Q8_0	7.83 GB	177.2 t/s	32k
OLMo 3 7B Instruct	7.3B	Runs great	Q8_0	9.72 GB	141.5 t/s	64k
Internlm3 8B Instruct	8.8B	Runs great	Q8_0	8.97 GB	154 t/s	32k
Mistral 7B Instruct V0.3	7.25B	Runs great	Q8_0	8.45 GB	164 t/s	32k
Gemma 4 E2B Instruct	5.12B	Runs great	Q8_0	5.44 GB	257.6 t/s	128k
Qwen3.5 4B	4.66B	Runs great	Q8_0	5.27 GB	268.6 t/s	256k
Mistral 7B Instruct V0.2	7.24B	Runs great	Q8_0	8.44 GB	164.1 t/s	32k
DeepSeek Coder 7B Instruct V1.5	6.91B	Runs great	Q8_0	10.58 GB	129.6 t/s	4k
DeepSeek Coder 6.7B Instruct	6.74B	Runs great	Q8_0	10.94 GB	125.1 t/s	16k
LFM2.5 8B A1B	8.47B	Runs great	Q8_0	8.81 GB	464.2 t/s	64k
Qwen3 4B	4.02B	Runs great	Q8_0	5.46 GB	258.4 t/s	32k
Gemma 3 4B Instruct	4.3B	Runs great	Q8_0	4.91 GB	289.7 t/s	128k
Phi 4 Mini Instruct	3.84B	Runs great	Q8_0	5.17 GB	274.7 t/s	128k
Granite 4.1 3B	3.4B	Runs great	Q8_0	4.34 GB	330.5 t/s	128k
Qwen2.5 3B Instruct	3.09B	Runs great	Q8_0	3.68 GB	394.9 t/s	32k
Phi 3 Mini 4k Instruct	3.82B	Runs great	Q8_0	4.8 GB	297.5 t/s	4k
Llama 3.2 3B Instruct	3.21B	Runs great	Q8_0	4.42 GB	325.4 t/s	128k
LFM2.5 2.6B	2.7B	Runs great	Q8_0	3.14 GB	471.4 t/s	128k
Qwen3.5 2B	2.27B	Runs great	Q8_0	2.7 GB	559.1 t/s	256k
Gemma 2 2B Instruct	2.61B	Runs great	Q8_0	3.54 GB	413 t/s	8k
Starcoder2 3B	3.03B	Runs great	Q8_0	3.34 GB	444.5 t/s	16k
Qwen3 1.7B	2.03B	Runs great	Q8_0	3.22 GB	457.2 t/s	32k
DeepSeek R1 Distill Qwen 1.5B	1.78B	Runs great	Q8_0	2.07 GB	753.8 t/s	128k
Qwen2.5 1.5B Instruct	1.54B	Runs great	Q8_0	2.06 GB	756.6 t/s	32k
SmolLM2 1.7B	1.71B	Runs great	Q8_0	3.53 GB	413 t/s	8k
OLMo 2 0425 1B	1.48B	Runs great	Q8_0	2.6 GB	583.1 t/s	4k
Llama 3.2 1B Instruct	1.24B	Runs great	Q8_0	1.82 GB	892.6 t/s	128k
LFM2.5 1.2B Instruct	1.17B	Runs great	Q8_0	1.46 GB	1178.6 t/s	64k
Gemma 3 1B Instruct	1B	Runs great	Q8_0	1.35 GB	1269.2 t/s	32k
Qwen3.5 0.8B	0.87B	Runs great	Q8_0	1.28 GB	1355.3 t/s	256k
Qwen3 0.6B	0.75B	Runs great	Q8_0	1.92 GB	815.3 t/s	32k
Qwen2.5 0.5B Instruct	0.49B	Runs great	Q8_0	0.88 GB	2278.6 t/s	32k
LFM2.5 350M	0.35B	Runs great	Q8_0	0.68 GB	3528.3 t/s	64k
Qwen3 Coder Next	79.67B	Partial offload	n/a	45.22 GB	88.9 t/s	n/a
Qwen3 Next 80B A3B Instruct	81.32B	Partial offload	n/a	46.15 GB	84.5 t/s	n/a
GPT OSS 120B	116.83B	Partial offload	n/a	59.4 GB	31.2 t/s	n/a
Qwen3.5 122B A10B	125.09B	Partial offload	n/a	70.61 GB	22 t/s	n/a
GLM 4.5 Air	110.47B	Partial offload	n/a	63.61 GB	13.1 t/s	n/a
MiniMax M2.5	228.7B	Partial offload	n/a	115.16 GB	10.7 t/s	n/a
MiniMax M2.7	228.69B	Partial offload	n/a	115.15 GB	10.7 t/s	n/a
Qwen3 235B A22B	235.09B	Partial offload	n/a	133.71 GB	4.5 t/s	n/a
MiniMax M3	427.04B	Partial offload	n/a	240.84 GB	4.4 t/s	n/a
GLM 4.5	358.34B	Partial offload	n/a	204.37 GB	3.3 t/s	n/a
DeepSeek V3.2	685.4B	Partial offload	n/a	385.91 GB	2.6 t/s	n/a
DeepSeek R1	684.53B	Partial offload	n/a	385.42 GB	2.6 t/s	n/a
Kimi K2.6	1026.88B	Partial offload	n/a	577.78 GB	2.5 t/s	n/a
Kimi K2 Instruct	1026.41B	Partial offload	n/a	577.52 GB	2.5 t/s	n/a
GLM 5.1	753.86B	Partial offload	n/a	424.51 GB	2.4 t/s	n/a
GLM 5.2	753.33B	Partial offload	n/a	424.21 GB	2.3 t/s	n/a
Llama 3 3 Nemotron Super 49B V1	49.87B	Partial offload	n/a	47.99 GB	1.7 t/s	n/a

What it will not run

0 of the 86 architectures we track are out of reach here, even at 2-bit precision. Another 17 run by splitting the model between the card and system RAM, which works but drops generation to single digits for most dense models.

If you want more headroom

The next steps up in memory, in order. More memory changes which models load at all; more bandwidth changes how fast they answer, so the two are worth weighing separately.

Device	Memory	Bandwidth	Models that fit
RTX 6000 Ada Generation	48 GB	960 GB/s	72	Check price
Radeon PRO W7900	48 GB	864 GB/s	72	Check price
RTX A6000	48 GB	768 GB/s	72	Check price
Ryzen AI Max+ 395 64GB	64 GB	256 GB/s	66	Check price

Price links go to an Amazon search for the model name and are affiliate links: if you buy through one, we earn a commission at no cost to you. We do not take payment for placement, and the ordering above is by memory capacity alone.