Yes — with 1.3 GB to spare
Mistral 7B Instruct v0.3 at Q4_K_M fits your GeForce RTX 4060 entirely on the GPU at 8K context, at an estimated 40 tokens per second. Past 18K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.
Fully on GPU
8K context
Q4_K_M · 4.1 GB
Apache 2.0
Released May 2024
Old but extremely well behaved, and permissively licensed for commercial use.
The VRAM budget
weights 4.1 GB
Weights 4.1 GB
KV cache @ 8K 1.00 GB
Runtime overhead 0.6 GB
Free 1.3 GB of 7.0 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| F16 | 13.5 GB | 15.1 GB | — | ~3.9 | Reference | 8.1 GB over |
| Q8_0 | 7.2 GB | 8.8 GB | — | ~12 | −0.1% ppl | 1.8 GB over |
| Q6_K | 5.5 GB | 7.1 GB | 6K | ~27 | −0.4% ppl | 0.1 GB over |
| Q5_K_M | 4.8 GB | 6.4 GB | 12K | 34 | −0.8% ppl | Fits |
| Q4_K_M | 4.1 GB | 5.7 GB | 18K | 40 | −1.9% ppl | Recommended |
| Q3_K_M | 3.3 GB | 4.9 GB | 24K | 50 | −5.4% ppl | Long context |
| Q2_K | 2.8 GB | 4.4 GB | 28K | 58 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ ollama pull mistral:7b $ OLLAMA_CONTEXT_LENGTH=8192 \ ollama run mistral:7b
The default. One binary, a model registry, an OpenAI-compatible port. More on Ollama.
01Download is 4.1 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03There is room to go to 18K context on this card.