What can AMD Radeon RX 9070 XT run for long context?

Build: RX 9070 XT + Ryzen 7 9700X + 32GB DDR5 (Linux/ROCm)

Memory: 16 GB VRAM + 32 GB system RAM
Runner: llama.cpp (ROCm)
AnyChatCodingAgentsReasoningVisionLong contextCreative

Runs comfortably
58 models

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

#1Phi-3.5 Mini Instruct
3.8B
phi
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 8.2 GBHeadroom: 7.8 GB
ollama run phi3.5:3.8b
95
tok/s
Estimated
Weights
4.10 GB
KV cache
1.90 GB
Activations
0.21 GB
Runtime
2.00 GB
Model details →
#2Codestral Mamba 7B
7B
mistral
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 6.1 GB
91
tok/s
Estimated
Weights
4.20 GB
KV cache
3.50 GB
Activations
0.22 GB
Runtime
2.00 GB
Model details →
#3Falcon Mamba 7B
7B
falcon
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 6.1 GB
91
tok/s
Estimated
Weights
4.20 GB
KV cache
3.50 GB
Activations
0.22 GB
Runtime
2.00 GB
Model details →
#4Gemma 4 E4B (Effective 4B)
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 8.6 GBHeadroom: 7.4 GB
ollama run gemma4:e4b
91
tok/s
Estimated
Weights
4.40 GB
KV cache
2.00 GB
Activations
0.23 GB
Runtime
2.00 GB
Model details →
#5Qwen 3 8B
8B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 11.0 GBHeadroom: 5.0 GB
ollama run qwen3:8b
80
tok/s
Estimated
Weights
4.80 GB
KV cache
4.00 GB
Activations
0.25 GB
Runtime
2.00 GB
Model details →
#6Command R7B (12-2024)
8B
command-r
Quant: Q4_K_MContext: 8,192VRAM: 10.6 GBHeadroom: 5.4 GB
80
tok/s
Estimated
Weights
4.40 GB
KV cache
4.00 GB
Activations
0.23 GB
Runtime
2.00 GB
Model details →
#7Ministral 8B Instruct
8B
mistral
Quant: Q4_K_MContext: 8,192VRAM: 11.3 GBHeadroom: 4.7 GB
80
tok/s
Estimated
Weights
5.00 GB
KV cache
4.00 GB
Activations
0.26 GB
Runtime
2.00 GB
Model details →
#8Qwen 2.5 7B Instruct
7B
qwen
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 11.0 GBHeadroom: 5.0 GB
ollama run qwen2.5:7b
52
tok/s
Estimated
Weights
8.10 GB
KV cache
0.47 GB
Activations
0.41 GB
Runtime
2.00 GB
Model details →
#9InternLM 2.5 7B Chat
7B
internlm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 10.1 GBHeadroom: 5.9 GB
91
tok/s
Estimated
Weights
4.40 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
2.00 GB
Model details →
#10Malhajar Mistral 7B Turkish
7.2B
mistral
Commercial OK
Quant: Q5_K_MContext: 8,192VRAM: 11.0 GBHeadroom: 5.0 GB
ollama run koezgen/malhajar-mistral-tur7b:latest
78
tok/s
Estimated
Weights
5.10 GB
KV cache
3.60 GB
Activations
0.26 GB
Runtime
2.00 GB
Model details →
#11Llama 3.2 3B Instruct
3B
llama
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 7.1 GBHeadroom: 8.9 GB
ollama run llama3.2:3b
121
tok/s
Estimated
Weights
3.40 GB
KV cache
1.50 GB
Activations
0.18 GB
Runtime
2.00 GB
Model details →
#12Phi-3.5 Vision
4.2B
phi
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 6.7 GBHeadroom: 9.3 GB
152
tok/s
Estimated
Weights
2.50 GB
KV cache
2.10 GB
Activations
0.13 GB
Runtime
2.00 GB
Model details →

Runs with tradeoffs
48 models

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

Mistral Nemo 12B Instruct
12B
mistral
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 15.9 GBHeadroom: 0.1 GB
  • • Tight VRAM fit — only 0.1 GB headroom left for context growth
ollama run mistral-nemo:12b
53
tok/s
Estimated
Weights
7.50 GB
KV cache
6.00 GB
Activations
0.38 GB
Runtime
2.00 GB
Model details →
Qwen 3 14B
14B
qwen
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 12.6 GBHeadroom: 3.4 GB
  • • Tight VRAM fit — only 3.4 GB headroom left for context growth
ollama run qwen3:14b
46
tok/s
Estimated
Weights
8.40 GB
KV cache
1.75 GB
Activations
0.42 GB
Runtime
2.00 GB
Model details →
Qwen 2.5 14B Instruct
14B
qwen
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 13.1 GBHeadroom: 2.9 GB
  • • Tight VRAM fit — only 2.9 GB headroom left for context growth
ollama run qwen2.5:14b
46
tok/s
Estimated
Weights
8.90 GB
KV cache
1.75 GB
Activations
0.45 GB
Runtime
2.00 GB
Model details →
Llama 3.1 8B Instruct
8B
llama
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 12.0 GBHeadroom: 4.0 GB
  • • Tight VRAM fit — only 4.0 GB headroom left for context growth
ollama run llama3.1:8b
45
tok/s
Estimated
Weights
8.50 GB
KV cache
1.07 GB
Activations
0.43 GB
Runtime
2.00 GB
Model details →
DeepSeek V2 Lite Chat
15.7B
deepseek
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 13.0 GBHeadroom: 3.0 GB
  • • Tight VRAM fit — only 3.0 GB headroom left for context growth
266
tok/s
Estimated
Weights
8.60 GB
KV cache
1.96 GB
Activations
0.43 GB
Runtime
2.00 GB
Model details →
Phi-4 14B
14B
phi
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 12.6 GBHeadroom: 3.4 GB
  • • Tight VRAM fit — only 3.4 GB headroom left for context growth
ollama run phi4:14b
46
tok/s
Estimated
Weights
8.40 GB
KV cache
1.75 GB
Activations
0.42 GB
Runtime
2.00 GB
Model details →
Yi Coder 9B
9B
yi
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 12.2 GBHeadroom: 3.8 GB
  • • Tight VRAM fit — only 3.8 GB headroom left for context growth
71
tok/s
Estimated
Weights
5.40 GB
KV cache
4.50 GB
Activations
0.28 GB
Runtime
2.00 GB
Model details →
GLM-4 9B
9B
glm
Quant: Q4_K_MContext: 8,192VRAM: 12.3 GBHeadroom: 3.7 GB
  • • Tight VRAM fit — only 3.7 GB headroom left for context growth
71
tok/s
Estimated
Weights
5.50 GB
KV cache
4.50 GB
Activations
0.28 GB
Runtime
2.00 GB
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: 28 new comfortable, 59 new tradeoff

  • • Qwen 3 0.6B
  • • Qwen 3 1.7B
  • • Gemma 3 270M
  • • SmolLM2 135M Instruct
Shop this upgrade↗

Upgrade to AMD Radeon RX 7900 XT

Launch MSRP $899 (2022). Check the current price.

20 GB VRAM (vs your 16 GB) plus a bandwidth jump from ~? GB/s to ~? GB/s.

Unlocks: 42 new comfortable

  • • Qwen 3 0.6B
  • • Llama 3.1 8B Instruct
  • • Qwen 3 1.7B
  • • Gemma 3 270M
Shop this upgrade↗

Add a second AMD Radeon RX 9070 XT

Launch MSRP $599 (2025). 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: 69 new comfortable

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

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

—
Qwen 3 235B-A22B
235B
qwen
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

—
DeepSeek R1 (671B reasoning)
671B
deepseek
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

—
Llama 4 Scout
109B
llama
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

—
DeepSeek V4 Flash (284B MoE)
284B
deepseek
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

—
Cloud GPU rental

Not enough VRAM for these models? Rent a GPU by the hour

The models above need more memory than 16 GB VRAM + 32 GB system RAM. You can rent a GPU with enough VRAM by the hour to run one of them, or to test it before you buy hardware. For daily use, owning a card usually costs less over time.

RunPod

billed per GPU-hour

Hourly GPU pods, in community-cloud and secure-cloud tiers.

The two tiers run on different hosts and terms. Check which tier a pod is on before a long job.

Rent on RunPod

Vast.ai

billed per GPU-hour

Marketplace of host-listed GPUs; price and reliability vary by host.

Host quality varies. Filter for verified hosts and check the host's reliability before a long run.

Rent on Vast.ai

Referral links: RunLocalAI may earn a commission or account credits if you sign up and rent, at no extra cost to you. Hourly rates and which GPUs are listed change all the time, so check the provider's page before you start. Editorial opinions are independent of the referral relationship; see how we make money.

How to read these numbers

Measured here
Measured here - RunLocalAI ran this exact combo on owner hardware with public evidence.

Source-backed
Source-backed / community - a reproduced public source supports the speed, but it is not labeled as owner-measured.

Extrapolated
Extrapolated - predicted from a measured benchmark on similar-bandwidth hardware.

Estimated
Estimated - formula based on VRAM bandwidth and model architecture; not a benchmark row.

RunLocalAI Will-It-Run Framework →

Want a specific benchmark we don't have? Email Contact support and we'll prioritize it.