Best AI Models for NVIDIA GeForce RTX 3090 (24 GB)

Runs 27–31B and 35B MoE models well; 70B doesn't fit in 24 GB.

24 GB GDDR6X936 GB/s71 TFLOPS FP16350 WAmpere 2020

What LLMs can the RTX 3090 run?

#	Model	Speed	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
2	Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
3	Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 29.8 t/s	29.8 tok/s	Q4_K_M · 21.7 GB
4	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB
5	NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA Q4_K_M · 19.3 GB · 191 t/s	191 tok/s	Q4_K_M · 19.3 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
2	Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 29.8 t/s	29.8 tok/s	Q4_K_M · 21.7 GB
3	Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 144 t/s	144 tok/s	Q4_K_M · 22.0 GB
4	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB
5	NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA Q4_K_M · 19.3 GB · 191 t/s	191 tok/s	Q4_K_M · 19.3 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
2	Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 29.8 t/s	29.8 tok/s	Q4_K_M · 21.7 GB
3	Qwen3.5 27BAlibaba Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
4	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB
5	NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA Q4_K_M · 19.3 GB · 191 t/s	191 tok/s	Q4_K_M · 19.3 GB

#	Model	Speed	Math Score	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s		Q4_K_M · 17.9 GB
2	Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s		Q4_K_M · 17.9 GB
3	Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 144 t/s	144 tok/s		Q4_K_M · 22.0 GB
4	Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 29.8 t/s	29.8 tok/s	overall	Q4_K_M · 21.7 GB
5	NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA Q4_K_M · 19.3 GB · 191 t/s	191 tok/s		Q4_K_M · 19.3 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
2	Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
3	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB
4	GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 179 t/s	179 tok/s	Q4_K_M · 13.1 GB
5	LFM2.5 24B A2BLiquid AI Q4_K_M · 15.0 GB · 213 t/s	213 tok/s	Q4_K_M · 15.0 GB

#	Model	Speed ↓	Quant · VRAM
1	Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 187 t/s	187 tok/s	Q4_K_M · 3.4 GB
2	NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA Q4_K_M · 19.3 GB · 191 t/s	191 tok/s	Q4_K_M · 19.3 GB
3	GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 179 t/s	179 tok/s	Q4_K_M · 13.1 GB
4	Qwen3 30B A3B Thinking 2507Alibaba Q4_K_M · 19.4 GB · 166 t/s	166 tok/s	Q4_K_M · 19.4 GB
5	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB

#	Model	Speed	Quant · VRAM
1	Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
2	Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB
3	Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 29.8 t/s	29.8 tok/s	Q4_K_M · 21.7 GB
4	Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 162 t/s	162 tok/s	Q4_K_M · 16.5 GB
5	JEV 27B VLautotrust Q4_K_M · 17.9 GB · 34.9 t/s	34.9 tok/s	Q4_K_M · 17.9 GB

See all 327 models on the RTX 3090 →Best quant that fits · speeds estimated on this card

About the RTX 3090

The RTX 3090 is a solid mid-range card in the 24 GB class: it runs 27–31B and 35B MoE models well. It is the most efficient per watt in its tier, and pricier per token than most of the cards compared at ≈$1,500 used. Ampere generation, released 2020.

Speed in 24 GB#3936 GB/s

Per watt#10.27 tok/s per W on Qwen3.5 9B (9B params)

Per dollar#56.4 tok/s per $100 on Qwen3.5 9B (9B params)

GenerationAmperereleased 2020

RTX 3090 strengths and weaknesses

Strengths

Near RTX 4090 speed: 93% of its speed for ≈$1,500 vs ≈$3,000 used, 350 W vs 450 W
Runs 35B MoE models: Qwen3.6 35B A3B at 144 tok/s in 22 GB
Room for long context: Qwen3.8 27B at Q4 uses 17.9 GB, leaving 6.1 GB free

Weaknesses

70B models don't fit: Llama 3.3 70B Instruct needs 51.3 GB at Q4 vs 24 GB here
Slow prompt reading: A 4K-token prompt on Qwen3.8 27B waits 7.1 s for the first token
No FP8 or NVFP4: Ampere predates both; use GGUF, AWQ or GPTQ builds
Used prices still near MSRP: ≈$1,500 used vs its $1,499 MSRP

Speed per dollar and per watt

NewUsed

On Qwen3.5 9B (9B params) at Q4_K_M. New prices are compared with new, used with used. The last two columns are the speed you get per $100 of price and per 100 W of board power; higher is better. The RTX 3090 has no new price, so the New view shows it at its MSRP.

Card	New price	tok/s	tok/s per $100	tok/s per 100 W
RX 7900 XTX24 GB	about $1,130	90.6	8.0	25.5
RTX 5070 Ti16 GBBest new value	about $1,170	91.6	7.8	30.5
RX 907016 GB	about $800	60.4	7.6	27.5
RX 9070 XT16 GB	about $815	60.4	7.4	19.9
RTX 309024 GB	about $1,499	95.7	6.4	27.3
RTX 508016 GB	about $1,690	98.1	5.8	27.3
RX 9060 XT 16GB16 GB	about $590	30.2	5.1	18.9
RTX 509032 GB	about $5,376	183	3.4	31.8

Card	Used price	tok/s	tok/s per $100	tok/s per 100 W
RX 907016 GBBest used value	about $650	60.4	9.3	27.5
RX 7900 XTX24 GB	about $1,000	90.6	9.1	25.5
RTX 5070 Ti16 GB	about $1,100	91.6	8.3	30.5
RTX 3090 Ti24 GB	about $1,600	103	6.4	22.9
RTX 309024 GB	about $1,500	95.7	6.4	27.3
RX 9060 XT 16GB16 GB	about $480	30.2	6.3	18.9
RTX 409024 GB	about $3,000	103	3.4	22.9
RTX 509032 GB	about $6,400	183	2.9	31.8

Similar GPUs

Cards people weigh against the RTX 3090, each with the one reason it is worth a look. Speeds on Qwen3.5 9B.

8% faster, same 24 GBRTX 3090 TiNVIDIA · 24 GB · Ampere

▲ 8% speedsame 24 GB$1,999rent $0.35/hr

About as fast, same 24 GBRX 7900 XTXAMD · 24 GB · RDNA 3

▼ 5% speedsame 24 GB$1,130

8% faster, same 24 GBRTX 4090NVIDIA · 24 GB · Ada Lovelace

▲ 8% speedsame 24 GB$1,599rent $0.43/hr

8 GB more and nearly 2× the speedRTX 5090NVIDIA · 32 GB · Blackwell

▲ 91% speed▲ 8 GB$5,376rent $0.50/hr

68% slower, 16 GBRX 9060 XT 16GBAMD · 16 GB · RDNA 4

▼ 68% speed▼ 8 GB$590

As fast, 16 GBRTX 5080NVIDIA · 16 GB · Blackwell

same speed▼ 8 GB$1,690rent $0.22/hr

RTX 3090 tokens per second by model size

Output speed depends on memory bandwidth (936 GB/s). Grey rows need more than 24 GB.

9BQwen3.5 9B · Q4_K_M · 6.6 GB
95.7tok/s
14BQwen3 14B · Q4_K_M · 11.9 GB
64.1tok/s
27–31BQwen3.8 27B · Q4_K_M · 17.9 GB
34.9tok/s
35B MoEQwen3.6 35B A3B · Q4_K_M · 22.0 GB
144tok/s
70BLlama 3.3 70B Instruct · not at 4-bit
needs 51.3 GB
120B MoEGPT OSS 120B · not at 4-bit
needs 70.8 GB

Estimated speeds: Q4_K_M, 4K context, one user. How we estimate

All 1141 models on the RTX 3090

Featured only

Best quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works

LLM models compatible with NVIDIA GeForce RTX 3090 — ranked by Score, recency and Fit
Model	Quant	VRAM	Speed	Context	Fit
MiniCPM V 2 68.1BVision Q4_K_M·113.7 t/s	Q4_K_M	7.9 GB33%	113.7 t/s	—	S94, Runs great
Devstral Small 250723.6BChat Q4_K_M·41.0 t/s	Q4_K_M	17.8 GB74%	41.0 t/s	131K	A75, Runs well
Gemma 3 1B Pt1000MChat Q4_K_M·922.0 t/s	Q4_K_M	2.1 GB9%	922.0 t/s	—	S100, Runs great
Qwen2.5 Coder 1.5B Instruct1.5BChatCodeFunctions Q4_K_M·475.4 t/s	Q4_K_M	1.7 GB7%	475.4 t/s	33K	S100, Runs great
Phi 3.5 Mini Instruct3.8BChatCode Q4_K_M·179.0 t/s	Q4_K_M	9.0 GB38%	179.0 t/s	131K	S99, Runs great
Llama Guard 3 1B1.5BChat Q4_K_M·614.7 t/s	Q4_K_M	2.6 GB11%	614.7 t/s	—	S100, Runs great
Granite 3.2 8B Instruct8.2BChatFunctions Q4_K_M·109.8 t/s	Q4_K_M	7.9 GB33%	109.8 t/s	131K	S93, Runs great
Mistral 7B v0.17.2BChat Q4_K_M·123.9 t/s	Q4_K_M	6.8 GB28%	123.9 t/s	33K	S95, Runs great
Qwen2.5 7B Instruct 1M7.6BChatFunctions Q4_K_M·121.9 t/s	Q4_K_M	5.8 GB24%	121.9 t/s	1010K	S95, Runs great
Qwen1.5 0.5B Chat620MChat Q4_K_M·699.5 t/s	Q4_K_M	2.3 GB9%	699.5 t/s	33K	S100, Runs great
Mellum2 12B A2.5B Instruct SFT12.1BChat Q4_K_M·245.9 t/s	Q4_K_M	7.8 GB32%	245.9 t/s	131K	S100, Runs great
Kumru 2B Base2.4BChat Q4_K_M·330.7 t/s	Q4_K_M	2.2 GB9%	330.7 t/s	8K	S100, Runs great
Qwen2.5 Coder 0.5B Instruct494MChatCodeFunctions Q4_K_M·981.5 t/s	Q4_K_M	0.8 GB3%	981.5 t/s	33K	S100, Runs great
Qwen3 Next 80B A3B Thinking81.3BChatFunctions IQ2_XXS·193.0 t/s	IQ2_XXS	22.9 GB95%	193.0 t/s	262K	B62, Decent
Tiny Starcoder Py164MChatCode Q4_K_M·5532.1 t/s	Q4_K_M	0.5 GB2%	5532.1 t/s	8K	S100, Runs great

RTX 3090 specs, power and PSU

Specs

Architecture: Ampere
VRAM: 24 GB GDDR6X
Memory bandwidth: 936.2 GB/s
FP16 (dense): 71.20 TFLOPS
CUDA cores: 10,496
Tensor cores: 328
Compute capability: 8.6
TDP: 350 W
Released: 2020-09-24
MSRP: $1,499

RTX 3090 FAQ

Is the RTX 3090 still worth it in 2026?

At ≈$1,500 used you get 6.4 tok/s per $100 on Qwen3.5 9B, which ranks 5th of the 8 cards compared here; it is also the 3rd fastest out of 5 cards with 24 GB. It runs 27–31B and 35B MoE models well; 70B doesn't fit in 24 GB.

What is the largest model the RTX 3090 can run?

Qwen3.6 35B A3B (36.0B) is the largest at 4-bit or better: at Q4_K_M it uses 22.0 GB of 24 GB, and runs at about 144 tok/s.

Can the RTX 3090 run Qwen3.8 27B?

Yes. Qwen3.8 27B fits at Q4_K_M, using 17.9 GB of 24 GB, and it is the best model that runs well on this card.

Can the RTX 3090 run 70B models?

No. Llama 3.3 70B Instruct needs 51.3 GB at Q4, more than the 24 GB on this card.

How many tokens per second does the RTX 3090 give?

About 34.9 tok/s on Qwen3.8 27B at Q4_K_M, with the first token after 7.1 s on a 4K-token prompt.

What power supply does the RTX 3090 need?

Plan a 700 W PSU (2× the 350 W TDP).

Is 24 GB enough for coding and agents?

Yes for coding: Qwen3.8 27B, the best of the 9 coding models that fit, runs at about 35 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.

Is renting cheaper than buying an RTX 3090?

Rent from $0.18/hr. Renting costs as much as buying after ≈8,200 hours (≈5.6 years at 4 h a day).

Best coding LLMs 24 GB VRAM 48 GB VRAM GPUs with 24 GB or more TITAN RTX RX 7900 XT