Skip to content
Run LLM Locally

Mistral 7B Instruct v0.3 on the RX 9070

With 16 GB of VRAM, the RX 9070 runs Mistral 7B Instruct v0.3 at Q4_K_M entirely on the GPU with an 8K context (about 6 GB needed). At Q8_0 it fits entirely.

ShareFacebookXWhatsAppLinkedInEmail

Mistral 7B Instruct v0.3

Q4_K_M · 8K context

Excellent

The model needs about 6 GB and fits in the 15.5 GB available, with 9.5 GB to spare.

Your hardware
RX 9070 16 GB · 64 GB RAM
Estimated memory
~6 GB estimate
VRAM (usable memory)
15.5 GB
GPU only
Yes
CPU/GPU offload
Not required
Generation speed
71–106 tok/s estimate
Largest context in fast memory
32K tokens
Confidence: low

Based on: memory bandwidth (640 GB/s) and model size

Memory breakdown

~6 GB (5.7 GB–6.4 GB) needed · estimate

Fits entirely in VRAM · 9.5 GB left

  • Model weights4.1 GB
  • KV cache1 GB
  • Runtime overhead0.6 GB
  • Safety margin0.3 GB
Assumptions: context 8K tokens · KV cache f16 · llama.cpp (ROCm) · flash attention on · weights = real file size

EstimateData updated 24 days agoSource: Run LLM Locally engine, Hugging Face config.json + model card

What makes this estimate less reliable

  • no benchmark for this hardware and model yet

Other quantizations on the RX 9070

Mistral 7B Instruct v0.3 on RX 9070 + 64 GB RAM · 8K context
Mistral 7B Instruct v0.3 Q8_08K~9.2 GB · 47–67 tok/s estimateExcellent
Mistral 7B Instruct v0.3 Q6_K8K~7.5 GB · 57–83 tok/s estimateExcellent
Mistral 7B Instruct v0.3 Q5_K_M8K~6.7 GB · 64–94 tok/s estimateExcellent
Mistral 7B Instruct v0.3 Q4_K_M8K~6 GB · 71–106 tok/s estimateExcellent
Mistral 7B Instruct v0.3 IQ4_XS8K~5.6 GB · 77–116 tok/s estimateExcellent
Mistral 7B Instruct v0.3 Q3_K_M8K~5.2 GB · 82–125 tok/s estimateExcellent

Alternatives

GPUs that run Mistral 7B Instruct v0.3 Q4_K_M entirely in VRAM

Other models that fit the RX 9070

Where to buy a RX 9070

Prices and links follow the market you choose.

Radeon RX 9070

16 GB GDDR6 · 640 GB/s · 220 W

ExcellentRuns Mistral 7B Instruct v0.3 Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

No current offer imported for this market: the link opens a marketplace search.

Frequently asked questions

Can the RX 9070 run Mistral 7B Instruct v0.3?
The model needs about 6 GB and fits in the 15.5 GB available, with 9.5 GB to spare.
Which quantization should I use on the RX 9070?
Q8_0 is the highest-precision quantization that fits entirely in VRAM at 8K context. ~9.2 GB / 15.5 GB.
What is the largest context on the RX 9070?
At Q4_K_M, about 32K tokens fit entirely in VRAM with an f16 KV cache. A q8_0 KV cache roughly doubles that.
How fast is Mistral 7B Instruct v0.3 (RX 9070)?
We estimate 71–106 tok/s (estimate, confidence low). Based on: memory bandwidth (640 GB/s) and model size.