What can NVIDIA RTX 2080 Ti 22GB (China-mod) run locally?

20 models fit comfortably on NVIDIA RTX 2080 Ti 22GB (China-mod) with 32 GB system RAM. 8 more run with tradeoffs. Top pick: Gemma 4 E2B (Effective 2B).

What's the recommended runner for this hardware?

llama.cpp / Ollama (CUDA). This is the runner that gives best tokens-per-second on NVIDIA RTX 2080 Ti 22GB (China-mod) based on windows support and the GPU backend.

What can NVIDIA RTX 2080 Ti 22GB (China-mod) run for vision?

Build: NVIDIA RTX 2080 Ti 22GB (China-mod) + — + 32 GB RAM (windows)

Memory: 22 GB VRAM + 32 GB system RAM

Runner: llama.cpp / Ollama (CUDA)

Any Chat Coding Agents Reasoning Vision Long context Creative

Runs comfortably
20 models

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

#1Gemma 4 E2B (Effective 2B)

gemma

Commercial OK

Quant: Q8_0Context: 8,192VRAM: 5.1 GBHeadroom: 16.9 GBTTFT: fast

ollama run gemma4:e2b

188

tok/s

Estimated

Weights

2.20 GB

KV cache

1.00 GB

Activations

0.12 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~190 ms (fast)

Model details →

#2Phi-3.5 Vision

4.2B

phi

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 6.5 GBHeadroom: 15.5 GBTTFT: fast

158

tok/s

Estimated

Weights

2.50 GB

KV cache

2.10 GB

Activations

0.13 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~400 ms (fast)

Model details →

#3Moondream 2

1.9B

other

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 3.3 GBHeadroom: 18.7 GBTTFT: fast

349

tok/s

Estimated

Weights

1.20 GB

KV cache

0.24 GB

Activations

0.06 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~181 ms (fast)

Model details →

#4Qwen 2.5-VL 3B

qwen

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 5.4 GBHeadroom: 16.6 GBTTFT: fast

221

tok/s

Estimated

Weights

2.00 GB

KV cache

1.50 GB

Activations

0.11 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~286 ms (fast)

Model details →

#5Qwen 2.5-VL 7B

qwen

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 12.1 GBTTFT: noticeable

tok/s

Estimated

Weights

4.40 GB

KV cache

3.50 GB

Activations

0.23 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~666 ms (noticeable)

Model details →

#6Qwen 2-VL 7B

qwen

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 10.1 GBHeadroom: 11.9 GBTTFT: noticeable

tok/s

Estimated

Weights

4.60 GB

KV cache

3.50 GB

Activations

0.24 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~666 ms (noticeable)

Model details →

#7LLaVA 1.6 Mistral 7B

other

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 10.0 GBHeadroom: 12.0 GBTTFT: noticeable

tok/s

Estimated

Weights

4.50 GB

KV cache

3.50 GB

Activations

0.23 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~666 ms (noticeable)

Model details →

#8LLaVA-OneVision 7B

other

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 10.0 GBHeadroom: 12.0 GBTTFT: noticeable

tok/s

Estimated

Weights

4.50 GB

KV cache

3.50 GB

Activations

0.23 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~666 ms (noticeable)

Model details →

#9Gemma 4 E4B (Effective 4B)

gemma

Commercial OK

Quant: Q8_0Context: 8,192VRAM: 8.4 GBHeadroom: 13.6 GBTTFT: fast

ollama run gemma4:e4b

tok/s

Estimated

Weights

4.40 GB

KV cache

2.00 GB

Activations

0.23 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~381 ms (fast)

Model details →

#10Gemma 3 4B

gemma

Commercial OK

Quant: Q8_0Context: 8,192VRAM: 8.4 GBHeadroom: 13.6 GBTTFT: fast

ollama run gemma3:4b

tok/s

Estimated

Weights

4.40 GB

KV cache

2.00 GB

Activations

0.23 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~381 ms (fast)

Model details →

#11MiniCPM-V 2.6 8B

minicpm

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 11.1 GBHeadroom: 10.9 GBTTFT: noticeable

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): ~761 ms (noticeable)

Model details →

#12MiniCPM-V 3 8B

minicpm

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 11.1 GBHeadroom: 10.9 GBTTFT: noticeable

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): ~761 ms (noticeable)

Model details →

Runs with tradeoffs
8 models

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

Phi-4 Multimodal

14B

phi

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 18.3 GBHeadroom: 3.7 GBTTFT: noticeable

• Tight VRAM fit — only 3.7 GB headroom left for context growth

tok/s

Estimated

Weights

9.00 GB

KV cache

7.00 GB

Activations

0.46 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~1332 ms (noticeable)

Model details →

Gemma 4 26B MoE

26B

gemma

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 21.9 GBHeadroom: 0.1 GBTTFT: slow

• Tight VRAM fit — only 0.1 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): ~2474 ms (slow)

Model details →

InternVL 2.5 26B

26B

other

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 21.9 GBHeadroom: 0.1 GBTTFT: slow

• Tight VRAM fit — only 0.1 GB headroom left for context growth

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): ~2474 ms (slow)

Model details →

Gemma 3 27B

27B

gemma

Commercial OK

Quant: Q4_K_MContext: 2,048VRAM: 22.0 GBHeadroom: 0.0 GBTTFT: slow

• Tight VRAM fit — only 0.0 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): ~2570 ms (slow)

Model details →

MedGemma 27B

27B

gemma

Quant: Q4_K_MContext: 2,048VRAM: 22.0 GBHeadroom: 0.0 GBTTFT: slow

• Tight VRAM fit — only 0.0 GB headroom left for context growth

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): ~2570 ms (slow)

Model details →

Qwen3.8 27B

27B

qwen

Commercial OK

Quant: Q5_K_MContext: 8,192VRAM: 36.1 GBHeadroom: 5.1 GBTTFT: slow

• Partial CPU offload: ~39% of layers run on CPU

ollama run qwen3.8:27b

tok/s

Estimated

Weights

19.80 GB

KV cache

13.50 GB

Activations

1.00 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~2570 ms (slow)

Model details →

Gemma 4 31B Dense

31B

gemma

Commercial OK

Quant: Q4_K_MContext: 8,192VRAM: 36.2 GBHeadroom: 5.0 GBTTFT: slow

• Partial CPU offload: ~39% 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): ~2950 ms (slow)

Model details →

PaliGemma 2 10B

10B

gemma

Commercial OK

Quant: BF16Context: 8,192VRAM: 27.8 GBHeadroom: 13.4 GBTTFT: noticeable

• Partial CPU offload: ~21% of layers run on CPU

tok/s

Estimated

Weights

20.00 GB

KV cache

5.00 GB

Activations

1.01 GB

Runtime

1.80 GB

Time to first token (prefill, 512-token prompt): ~952 ms (noticeable)

Model details →

What if you upgraded?

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

+32 GB system RAM

Check the current price

Adds 32 GB to your CPU-offload working set. Helps when models don't quite fit in VRAM.

Unlocks: 96 new comfortable, 62 new tradeoff

• Qwen 3 0.6B
• Llama 3.1 8B Instruct
• Qwen 3 8B
• Qwen 3 1.7B

Shop this upgrade↗

Upgrade to NVIDIA GeForce RTX 3090 Ti

Launch MSRP $1,999 (2022). Now sells well below its launch price on the used market. Check the current price.

24 GB VRAM (vs your 22 GB) plus a bandwidth jump from ~616 GB/s to ~? GB/s.

Unlocks: 111 new comfortable

• Qwen 3 0.6B
• Llama 3.1 8B Instruct
• Qwen 3 8B
• Qwen 3 1.7B

Shop this upgrade↗

Add a second NVIDIA RTX 2080 Ti 22GB (China-mod)

Check the current price

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: 139 new comfortable

• Qwen 3 0.6B
• Llama 3.1 8B Instruct
• Qwen 3 30B-A3B
• Qwen 2.5 Coder 32B Instruct

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