Run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF locally
MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF is a mid-size instruction-tuned chat model with 7.25 billion parameters, built on the llama architecture. It is released under the apache-2.0 license and has been downloaded 180,190 times.
To run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF locally at a 4,096-token context, its quantized versions need between 2.8 GB (IQ1_S, lowest quality) and 14.8 GB (GGUF, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is Q2_K, needing about 3.84 GB. That means MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.
All quantizations
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| IQ1_S | 1.78 | Very low | 1.5 GB | 0.5 GB | 2.8 GB | 265.9 t/s | Fits in VRAM |
| IQ1_M | 1.94 | Very low | 1.64 GB | 0.5 GB | 2.94 GB | 244.4 t/s | Fits in VRAM |
| IQ2_XS | 2.43 | Very low | 2.05 GB | 0.5 GB | 3.35 GB | 195.1 t/s | Fits in VRAM |
| Q2_K | 3.01 | Low | 2.54 GB | 0.5 GB | 3.84 GB | 157.7 t/s | Fits in VRAM |
| IQ3_XS | 3.34 | Fair | 2.82 GB | 0.5 GB | 4.12 GB | 17.8 t/s | Offload |
| Q3_K_S | 3.5 | Fair | 2.95 GB | 0.5 GB | 4.25 GB | 16.9 t/s | Offload |
| Q3_K_M | 3.89 | Fair | 3.28 GB | 0.5 GB | 4.58 GB | 15.2 t/s | Offload |
| Q3_K_L | 4.22 | Good | 3.56 GB | 0.5 GB | 4.86 GB | 14.0 t/s | Offload |
| IQ4_XS | 4.32 | Good | 3.64 GB | 0.5 GB | 4.94 GB | 13.7 t/s | Offload |
| Q4_K_S | 4.57 | Good | 3.86 GB | 0.5 GB | 5.16 GB | 13.0 t/s | Offload |
| Q4_K_M | 4.83 | Good | 4.07 GB | 0.5 GB | 5.37 GB | 12.3 t/s | Offload |
| Q5_K_S | 5.52 | Very good | 4.66 GB | 0.5 GB | 5.96 GB | 10.7 t/s | Offload |
| Q5_K_M | 5.67 | Very good | 4.78 GB | 0.5 GB | 6.08 GB | 10.5 t/s | Offload |
| Q6_K | 6.56 | Excellent | 5.54 GB | 0.5 GB | 6.84 GB | 9.0 t/s | Offload |
| Q8_0 | 8.5 | Excellent | 7.17 GB | 0.5 GB | 8.47 GB | 7.0 t/s | Offload |
| GGUF | 16.0 | Excellent | 13.5 GB | 0.5 GB | 14.8 GB | 3.7 t/s | Offload |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
How much VRAM do you need to run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF?
You need about 5.96 GB of VRAM to run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF entirely on the GPU using the Q5_K_S quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF on an 8 GB GPU?
Yes. With 8 GB of VRAM you can run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF fully on the GPU using Q6_K (about 6.84 GB).
Can I run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF fully on the GPU using GGUF (about 14.8 GB).
Can I run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF fully on the GPU using GGUF (about 14.8 GB).
What is the best quantization for MaziyarPanahi/Mistral-7B-Instruct-v0.3-GGUF?
If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.