What can NVIDIA GeForce RTX 4090 run locally?

175 models fit comfortably on NVIDIA GeForce RTX 4090 with 32 GB system RAM. 56 more run with tradeoffs. Top pick: Qwen 3 0.6B.

What's the recommended runner for this hardware?

llama.cpp / Ollama (CUDA). This is the runner that gives best tokens-per-second on NVIDIA GeForce RTX 4090 based on windows support and the GPU backend.

What can NVIDIA GeForce RTX 4090 run for chat?

Build: NVIDIA GeForce RTX 4090 + — + 32 GB RAM (windows)

Memory: 24 GB VRAM + 32 GB system RAM

Runner: llama.cpp / Ollama (CUDA)

Any Chat Coding Agents Reasoning Vision Long context Creative

Runs comfortably
175 models

Ranked by fit for chat use case + predicted speed. Click a row for VRAM breakdown.

#1Qwen 3 0.6B

0.6B

qwen

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 2.4 GBHeadroom: 21.6 GBTTFT: instant

1809

tok/s

Estimated

Weights

0.30 GB

KV cache

0.30 GB

Activations

0.02 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~19 ms (instant)

Model details →

#2Llama 3.2 3B Instruct

llama

Commercial OK

Quant: Q8_0Context: 8,192VRAM: 6.9 GBHeadroom: 17.1 GBTTFT: instant

ollama run llama3.2:3b

206

tok/s

Estimated

Weights

3.40 GB

KV cache

1.50 GB

Activations

0.18 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~93 ms (instant)

Model details →

#3Qwen 3 1.7B

1.7B

qwen

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 3.6 GBHeadroom: 20.4 GBTTFT: instant

638

tok/s

Estimated

Weights

0.90 GB

KV cache

0.85 GB

Activations

0.05 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~53 ms (instant)

Model details →

#4Dolphin 3.0 8B

dolphin

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 11.1 GBHeadroom: 12.9 GBTTFT: fast

ollama run dolphin3:8b

136

tok/s

Estimated

Weights

5.00 GB

KV cache

4.00 GB

Activations

0.26 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~248 ms (fast)

Model details →

#5TinyLlama 1.1B Chat v1.0

1.1B

llama

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 2.6 GBHeadroom: 21.4 GBTTFT: instant

987

tok/s

Estimated

Weights

0.60 GB

KV cache

0.14 GB

Activations

0.03 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~34 ms (instant)

Model details →

#6Gemma 2 2B Instruct

gemma

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 4.0 GBHeadroom: 20.0 GBTTFT: instant

543

tok/s

Estimated

Weights

1.10 GB

KV cache

1.00 GB

Activations

0.06 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~62 ms (instant)

Model details →

#7Mistral 7B Instruct v0.3

mistral

Commercial OK

Quant: Q5_K_MContext: 8,192VRAM: 10.7 GBHeadroom: 13.3 GBTTFT: fast

ollama run mistral:7b

136

tok/s

Estimated

Weights

5.10 GB

KV cache

3.50 GB

Activations

0.26 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~217 ms (fast)

Model details →

#8DeepSeek V2 Lite Chat

15.7B

deepseek

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 18.7 GBHeadroom: 5.3 GBTTFT: instant

452

tok/s

Estimated

Weights

8.60 GB

KV cache

7.85 GB

Activations

0.44 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~74 ms (instant)

Model details →

#9Trendyol LLM 7B Chat v0.1

llama

Quant: Q4_K_MContext: 4,096VRAM: 7.6 GBHeadroom: 16.4 GBTTFT: fast

155

tok/s

Estimated

Weights

3.90 GB

KV cache

1.75 GB

Activations

0.20 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~217 ms (fast)

Model details →

#10Kumru 2B

2.4B

mistral

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 4.6 GBHeadroom: 19.4 GBTTFT: instant

ollama run alibayram/kumru:latest

452

tok/s

Estimated

Weights

1.50 GB

KV cache

1.20 GB

Activations

0.08 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~74 ms (instant)

Model details →

#11Salamandra 7B Instruct

llama

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 9.4 GBHeadroom: 14.6 GBTTFT: fast

155

tok/s

Estimated

Weights

3.90 GB

KV cache

3.50 GB

Activations

0.20 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~217 ms (fast)

Model details →

#12TinyLlama 1.1B Chat v0.3 AWQ

1.1B

other

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 2.6 GBHeadroom: 21.4 GBTTFT: instant

987

tok/s

Estimated

Weights

0.60 GB

KV cache

0.14 GB

Activations

0.03 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~34 ms (instant)

Model details →

Runs with tradeoffs
56 models

Tight VRAM, partial CPU offload, or context-limited.

GPT-OSS Swallow 20B RL v0.1

20B

other

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 23.4 GBHeadroom: 0.6 GBTTFT: noticeable

• Tight VRAM fit — only 0.6 GB headroom left for context growth

tok/s

Estimated

Weights

11.00 GB

KV cache

10.00 GB

Activations

0.56 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~620 ms (noticeable)

Model details →

Sarvam M

24B

mistral

Commercial OK

Quant: Q4_K_MContext: 4,096VRAM: 21.7 GBHeadroom: 2.3 GBTTFT: noticeable

• Tight VRAM fit — only 2.3 GB headroom left for context growth

tok/s

Estimated

Weights

13.20 GB

KV cache

6.00 GB

Activations

0.66 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~744 ms (noticeable)

Model details →

Gemma 4 26B MoE

26B

gemma

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 21.9 GBHeadroom: 2.1 GBTTFT: noticeable

• Tight VRAM fit — only 2.1 GB headroom left for context growth

ollama run gemma4:26b-moe

tok/s

Estimated

Weights

16.00 GB

KV cache

3.25 GB

Activations

0.80 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~806 ms (noticeable)

Model details →

Gemma 3 27B

27B

gemma

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 22.0 GBHeadroom: 2.0 GBTTFT: noticeable

• Tight VRAM fit — only 2.0 GB headroom left for context growth

ollama run gemma3:27b

tok/s

Estimated

Weights

16.00 GB

KV cache

3.38 GB

Activations

0.80 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~837 ms (noticeable)

Model details →

Muse Glimmer 30B

30B

other

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 20.4 GBHeadroom: 3.6 GBTTFT: noticeable

• Tight VRAM fit — only 3.6 GB headroom left for context growth

ollama run muse-glimmer

tok/s

Estimated

Weights

17.00 GB

KV cache

0.71 GB

Activations

0.86 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~930 ms (noticeable)

Model details →

Gemma 4 31B Dense

31B

gemma

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 36.2 GBHeadroom: 7.0 GBTTFT: noticeable

• Partial CPU offload: ~34% of layers run on CPU

ollama run gemma4:31b

tok/s

Estimated

Weights

18.00 GB

KV cache

15.50 GB

Activations

0.91 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~961 ms (noticeable)

Model details →

OLMo 2 32B

32B

other

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 37.8 GBHeadroom: 5.4 GBTTFT: noticeable

• Partial CPU offload: ~36% of layers run on CPU

tok/s

Estimated

Weights

19.00 GB

KV cache

16.00 GB

Activations

0.96 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~992 ms (noticeable)

Model details →

EXAONE 3.5 32B Instruct AWQ

32B

other

Quant: Q4_K_MContext: 8,192VRAM: 36.3 GBHeadroom: 6.9 GBTTFT: noticeable

• Partial CPU offload: ~34% of layers run on CPU

tok/s

Estimated

Weights

17.60 GB

KV cache

16.00 GB

Activations

0.89 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~992 ms (noticeable)

Model details →

What if you upgraded?

Hypothetical scenarios. We re-ran the compatibility engine for each.

+32 GB system RAM

~$80–150

Doubles your CPU-offload working set. Helps when models don't quite fit in VRAM.

Unlocks: 9 new comfortable, 73 new tradeoff

• Gemma 3 270M
• SmolLM2 135M Instruct
• SmolLM2 360M Instruct
• VBART Large (Turkish Summarization)

Shop this upgrade↗

Upgrade to NVIDIA RTX PRO 4500 Blackwell

see current pricing

32 GB VRAM (vs your 24 GB) plus a bandwidth jump from ~1008 GB/s to ~896 GB/s.

Unlocks: 48 new comfortable

• Qwen 3 30B-A3B
• Qwen 2.5 Coder 32B Instruct
• Qwen3.6 27B
• Qwen 3 32B

Shop this upgrade↗

Add a second NVIDIA GeForce RTX 4090

~$1899

Tensor parallelism splits the model across both cards, effectively doubling VRAM. Bandwidth doesn't double — runs ~1.5× the single-card speed in practice.

Unlocks: 61 new comfortable

• Qwen 3 30B-A3B
• Qwen 2.5 Coder 32B Instruct
• Qwen3.6 27B
• Qwen 3 32B

Shop this upgrade↗

Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.

Won't run
top 5 popular models

Need more memory than you have. Shown for orientation.

DeepSeek V4 Pro (1.6T MoE)

1600B

deepseek

Commercial OK