Mistral-7B-Instruct-v0.3 on RTX 4060: Fits (Sep 2026)
Conditional memory judgement for one configuration. Change context, quant, or engine and the conclusion can change. Configuration used on this page: GGUF Q4_K_M (bartowski), 8,192 cached tokens, 1 sequence, engine llama.cpp. Status: Likely fits. Software: Not verified..
Verdict on GeForce RTX 4060 8GB (desktop)
Yes, run Q4_K_M. The calculator puts it at 5.2 GiB–5.7 GiB at 8,192 tokens, inside 7.0 GiB of usable VRAM (after the driver reserve). Q8_0 is the next step up and does not clear the bar (8.3 GiB–8.9 GiB).
With Q4_K_M, the longest context that still counts as a fit is about 12,602 tokens of a 32,768-token native window, one sequence, fp16 KV cache.
| GGUF quant | File weights | KV cache | Estimated total | Verdict |
|---|---|---|---|---|
| Q8_0 | 7.2 GiB | 1.0 GiB | 8.3 GiB–8.9 GiB | Does not fit |
| Q6_K | 5.5 GiB | 1.0 GiB | 6.7 GiB–7.2 GiB | Tight |
| Q5_K_M | 4.8 GiB | 1.0 GiB | 5.9 GiB–6.4 GiB | Tight |
| Q4_K_M (recommended) | 4.1 GiB | 1.0 GiB | 5.2 GiB–5.7 GiB | Fits |
Engine llama.cpp, one sequence, fp16 KV. File weights are the published GGUF sizes in the catalog. Totals come from the same calculator that powers the tool below; they are estimates, not a measurement of this setup.
Fixed page configuration
GGUF Q4_K_M needs 5.2 GiB-5.7 GiB at 8,192 tokens against 7.00 GiB net on GeForce RTX 4060 8GB (desktop). Q8_0, Q6_K, Q5_K_M are tight or over budget.
- ModelMistral-7B-Instruct-v0.3
- HardwareGeForce RTX 4060 8GB (desktop)
- QuantizationGGUF Q4_K_M (bartowski)
- Context length (cached tokens)8,192
- Engine
llama.cpp - Concurrent sequences1
- KV cache typefp16
Technical identifiers and share data
- Model
mistral-7b-instruct-v0.3 - Hardware
nvidia-geforce-rtx-4060-8gb-desktop - Quantization
gguf-q4_k_m - Share fragment for this identity
v2.eyJzY2hlbWFWZXJzaW9uIjoyLCJjYWxjdWxhdG9yVmVyc2lvbiI6IjAuMy4wIiwiZGF0YXNldFZlcnNpb24iOiIyMDI2LTEwLTAyLjEiLCJjb25maWciOnsibW9kZWxJZCI6Im1pc3RyYWwtN2ItaW5zdHJ1Y3QtdjAuMyIsImhhcmR3YXJlSWQiOiJudmlkaWEtZ2Vmb3JjZS1ydHgtNDA2MC04Z2ItZGVza3RvcCIsImRldmljZUNvdW50IjoxLCJxdWFudGl6YXRpb25JZCI6ImdndWYtcTRfa19tIiwiY29udGV4dFRva2VucyI6ODE5MiwiY29uY3VycmVudFNlcXVlbmNlcyI6MSwiZW5naW5lIjoibGxhbWEuY3BwIiwia3ZUeXBlIjoiZnAxNiJ9fQ
Memory breakdown
- Weight artifact
- 4.07 GiB (4,372,812,000 bytes)
- Raw KV cache
- 1.00 GiB (1,073,741,824 bytes)
- Engine extras / workspace
- 0.13 GiB–0.65 GiB
- Total estimated demand
- 5.20 GiB–5.72 GiB
- Net calculator budget
- 7.00 GiB
- OS / display reserve
- 1.00 GiB
- Remaining vs net budget
- 1.28 GiB–1.80 GiB
5.72 GiB high estimate vs 7.00 GiB net
Enough memory does not mean this engine, quant, and OS will run. Support does not mean it will be fast.
Other published files on this device
Same model, device and context (8,192 tokens, 1 sequence, FP16 KV). Each row is one published weight-file set with its usual engine, sorted by memory.
| Quantization | Engine | Estimated memory | Fit |
|---|---|---|---|
| BF16 safetensors | vllm | 15.13 GiB–17.03 GiB | Does not fit |
| GGUF Q8_0 (bartowski) | llama.cpp | 8.30 GiB–8.89 GiB | Does not fit |
| GGUF Q6_K (bartowski) | llama.cpp | 6.66 GiB–7.22 GiB | Tight fit |
| GGUF Q5_K_M (bartowski) | llama.cpp | 5.91 GiB–6.45 GiB | Tight fit |
| GGUF Q4_K_M (bartowski) This page | llama.cpp | 5.20 GiB–5.72 GiB | Fits |
Assumptions that affect this result
- Requested context is within the recorded native window of 32768 tokens.
- Full-resident weight estimates use total parameters, including inactive MoE experts.
- Weight bytes come from the selected weight files only, not from every file in the repository.
- Download file size is not peak GPU memory. Runtime layout, allocator padding, and KV are extra.
- Counted files: Mistral-7B-Instruct-v0.3-Q4_K_M.gguf.
- Context budget is the number of cached tokens per sequence (prompt plus reserved generation).
- Concurrent sequences share one model replica; weights are not multiplied by sequence count.
- KV dtype is independent of weight quantization unless you change it.
Related pages
Sources
- Mistral-7B-Instruct-v0.3 model card
- config.json — revision c170c708c41dac9275d15a8fff4eca08d52bab71
- safetensors index — metadata.total_size=14496047104
- Artifact repo bartowski/Mistral-7B-Instruct-v0.3-GGUF
- NVIDIA GeForce RTX 40 Series specs — Memory Size 8 GB GDDR6
Calculator
This combo page keeps the route model and device. Customize to change them on the main calculator; the share then keeps that identity.
Loading catalog snapshot…
Search or pick a catalog checkpoint to see memory demand.