# Model: Ministral 3 14B Instruct > Source: > Published: 2026-11-14 12:00:00+00:00 # Ministral 3 14B Instruct 14B Mistral Ministral 3 The larger Ministral 3 instruction model under Apache 2.0, with text and image input. Its language backbone has 40 full-attention layers, 8 KV heads of 128 dimensions and a 256K window. Parameters count the language model reconstructed from the official config; its separate vision encoder is recorded as an additional download, not folded into the language weights. Long-context and image use need more memory than the GGUF file alone. No per-level perplexity loss has been published. 262K Max Context 3 Quant Variants GGUF Q8_0 Best Quality — Perplexity rise (best build) ## Quantization Variants Per-quant VRAM, quality loss, and inference speed on RTX 4090 Measured = run on this site · Estimated = not run here: calculated, or a figure this site has not verified · Community = public reports | Format | Level | BPW | VRAM | PPL Loss | Speed | Source | Actions | |---|---|---|---|---|---|---|---| | GGUF | Q4_K_M | 4.88 | 9.3 GB | — | — | Estimated | | | GGUF | Q5_K_M | 5.7 | 10.7 GB | — | — | Estimated | | | GGUF | Q8_0 | 8.51 | 15.7 GB | — | — | Estimated | | ### Which file each size is for A file size is not a memory figure: the VRAM column is the calculator's total at 4K context — weights plus KV cache plus a 10% activation buffer. Units differ on purpose: file sizes are decimal GB, as Hugging Face lists them; the VRAM column is GiB (1024³ bytes) with 2% added to the weights, so 7.12 GB of file counts as about 6.8 GiB there. A vision projector (mmproj), where the repo ships one, is a separate download and is in neither number. - GGUF Q4_K_M[mistralai/Ministral-3-14B-Instruct-2512-GGUF](https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512-GGUF) · Ministral-3-14B-Instruct-2512-Q4_K_M.gguf · 8.24 GB (8,239,593,024 bytes) · + vision projector 0.88 GB, separate · listing read 2026-11-14 - GGUF Q5_K_M[mistralai/Ministral-3-14B-Instruct-2512-GGUF](https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512-GGUF) · Ministral-3-14B-Instruct-2512-Q5_K_M.gguf · 9.62 GB (9,621,091,904 bytes) · + vision projector 0.88 GB, separate · listing read 2026-11-14 - GGUF Q8_0[mistralai/Ministral-3-14B-Instruct-2512-GGUF](https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512-GGUF) · Ministral-3-14B-Instruct-2512-Q8_0.gguf · 14.36 GB (14,359,836,224 bytes) · + vision projector 0.88 GB, separate · listing read 2026-11-14 ## Similar models [Compare with Ministral 3](https://quantized.uk/tools/compare/?a=ministral-3-14b&b=ministral-3-3b) [3BConsumer GPUMac / Apple SiliconQ4_K_M2.7 GB—perplexity rise](https://quantized.uk/quant-hub/ministral-3-3b/) ### Ministral 3 3B Instruct Mistral Ministral 3 The small Ministral 3 instruction model under Apache 2.0, with text and image input. The language backbone has 26 full-attention layers and a 256K window. The language parameter count is reconstructed from the official config; the separate vision encoder is excluded from that count and its bytes are recorded beside each GGUF. Image-input sizing must include that encoder and budget image tokens within the context. No per-level perplexity loss has been published. [8BConsumer GPUMac / Apple SiliconQ4_K_M6.0 GB—perplexity rise](https://quantized.uk/quant-hub/ministral-3-8b/) ### Ministral 3 8B Instruct Mistral Ministral 3 Mistral ships the GGUF itself, in 3B / 8B / 14B and Instruct or Reasoning variants — the 8B is the one that fits an 8GB card at Q4 with room for real context. Plain full attention on all 34 layers, so the memory estimates below behave the way the calculator assumes. Quality loss per level has not been published. [8BConsumer GPUMac / Apple SiliconQ4_K_M6.2 GB+3.0%perplexity rise](https://quantized.uk/quant-hub/qwen3-vl-8b/) ### Qwen3-VL 8B Instruct Alibaba Qwen3-VL Current-generation vision-language model that still fits a single 8–12GB card at Q4 (~5.9GB). The realistic multimodal option for people without a 24GB GPU — note the vision encoder adds VRAM the KV-cache math below does not model. [9BConsumer GPUMac / Apple SiliconQ4_K_M6.6 GB—perplexity rise](https://quantized.uk/quant-hub/qwen3-5-9b/) ### Qwen3.5 9B Alibaba Qwen3.5 The newest small Qwen and a common pick for 8 to 16 GB cards. Only 8 of its 32 layers keep a growing KV cache, so it can use a long context on modest hardware. At Q4_K_M it needs about 6.6 GB at 4K and 7.0 GB at 16K, comfortable on an 8 GB card; at 32K (7.5 GB) it is tight there and comfortable from 10 GB up. Its full 256K window needs about 15.2 GB. Q8_0 needs about 10.4 GB, comfortable from 12 GB up. Reasoning is off by default on the small Qwen3.5 sizes, and it reads images through a separate vision encoder. Sizes come from the published GGUF file sizes, and no per-level quality loss has been published. ## Running Ministral 3 14B Instruct locally At Q4_K_M and 4K of context, Ministral 3 14B Instruct needs about 9.3 GB — 7.8 GB of weights, 0.63 GB of KV cache and a 0.8 GB activation buffer. The smallest cards in this index that clear that comfortably are 12 GB ones — 10 of them, from the RTX 3060 12G up — and 65 of the 78 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs. ## What longer context costs Going from 4K to 32K adds about 4.37 GB, taking the total to 14.1 GB. Weights do not move with context — only the KV cache does, and it grows linearly, so this is the number to watch when planning for long documents. The model's native window is 256K; holding all of it at Q4_K_M would need about 53 GB. ## Which build to download This index only tracks Ministral 3 14B Instruct in GGUF, across 3 levels (Q4_K_M, Q5_K_M, Q8_0). No per-level perplexity sweep has been published for this model, so the quality column is empty rather than estimated — within one model, more bits per weight is the only ordering the data supports. No throughput has been measured for this model on this site's hardware. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 — CUDA only, and now archived, so local use rather than serving. ## Common questions - How much VRAM does Ministral 3 14B Instruct need? - About 9.3 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 14.1 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card. - Will Ministral 3 14B Instruct run on a 12GB GPU? - Yes — at Q4_K_M and 4K context it needs about 9.3 GB, which leaves 2.7 GB spare on a RTX 3060 12G, or on any of the other 12 GB cards here (10 in all). That is the smallest size in this index that clears it comfortably; 65 of 78 do. - Which quantization of Ministral 3 14B Instruct should I use? - No published quality comparison exists for this model, so pick by footprint: Q4_K_M is the level most people run, and a higher bits-per-weight level is more faithful. All 3 levels in the index are Q4_K_M, Q5_K_M, Q8_0. ## Where this model fits Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory. - Runs comfortably on - [RTX 3060 12G12GB · +2.7](https://quantized.uk/gpu/rtx-3060-12g/)[Radeon RX 6700 XT12GB · +2.7](https://quantized.uk/gpu/radeon-rx-6700-xt/)[Radeon RX 7700 XT12GB · +2.7](https://quantized.uk/gpu/radeon-rx-7700-xt/)[Arc B580 12G12GB · +2.7](https://quantized.uk/gpu/arc-b580-12g/)[RTX 4070 Ti12GB · +2.7](https://quantized.uk/gpu/rtx-4070-ti/) - Tight but possible - [Arc B570 10G10GB](https://quantized.uk/gpu/arc-b570-10g/)[RTX 3080 10G10GB](https://quantized.uk/gpu/rtx-3080-10g/) - Just misses - [RTX 4060+1.3GB over](https://quantized.uk/gpu/rtx-4060/)[RTX 4060 Ti 8G+1.3GB over](https://quantized.uk/gpu/rtx-4060-ti-8g/) - Ships in - [GGUF](https://quantized.uk/formats/gguf/) - A top pick for - [Best local LLM for 12GB](https://quantized.uk/best/12gb-vram/) - Try it yourself - [Size it in the calculator](https://quantized.uk/tools/vram-calc/?model=ministral-3-14b&quant=Q4_K_M)[Generate the run command](https://quantized.uk/tools/cli-gen/?model=ministral-3-14b&quant=Q4_K_M) ## How to actually run this Deployment guides for this model and this class of hardware. This page's figures change when the model or the runtime does.Last updated 2026-11-14 [RSS → /feed.xml](https://quantized.uk/feed.xml)