| --- |
| library_name: mlx |
| language: |
| - en |
| - fr |
| - es |
| - de |
| - it |
| - pt |
| - nl |
| - zh |
| - ja |
| - ko |
| - ar |
| license: apache-2.0 |
| inference: false |
| extra_gated_description: If you want to learn more about how we process your personal |
| data, please read our <a href="https://mistral.ai/terms/">Privacy Policy</a>. |
| tags: |
| - mistral-common |
| - mlx |
| - ministral |
| - ministral-3 |
| - vision-language |
| - multimodal |
| - quantized |
| - edge |
| - 6-bit |
| - base-model |
| pipeline_tag: image-text-to-text |
| base_model: mistralai/Ministral-3-8B-Base-2512 |
| --- |
| |
| # mlx-community/Ministral-3-8B-Base-2512-6bit |
|
|
| This is, **Ministral 3 8B Base 2512** is a vision-language model: a text backbone paired |
| with a vision encoder, supporting image understanding alongside text. This is |
| the **base pre-trained** checkpoint — not instruction- or chat-tuned. For |
| chat/instruction-following use cases, use the |
| [Instruct variant](https://huggingface.co/mlx-community/Ministral-3-8B-Instruct-2512-6bit) |
| instead; this base checkpoint is intended for custom post-training/fine-tuning. |
|
|
| > **Community note.** Structural check confirms the vision tower and |
| > multimodal projector were carried over intact (not dropped, which is a real |
| > failure mode for text-only conversion tools on vision-language models). |
| > Functional check confirms both text-only and image+text generation produce |
| > coherent output. Converted and verified by a single maintainer running |
| > local MLX tooling -- not independently reviewed by anyone else; please open |
| > a discussion if you hit anything unexpected. |
|
|
| This is an MLX conversion of [`mistralai/Ministral-3-8B-Base-2512`](https://huggingface.co/mistralai/Ministral-3-8B-Base-2512), |
| converted with [mlx-vlm](https://github.com/Blaizzy/mlx-vlm). Refer to the |
| [original model card](https://huggingface.co/mistralai/Ministral-3-8B-Base-2512) for the full |
| description, capabilities, and license terms. |
|
|
| ## Heads up |
|
|
| - **Base model, not instruct-tuned** — expect raw completion behavior, not |
| chat-following. Don't expect it to follow instructions well. |
| - **Vision retained at full precision** — only the language backbone is |
| quantized; the vision tower and multimodal projector are untouched bf16, |
| per mlx-vlm's standard policy of not quantizing multimodal modules. |
| - **Output size on disk: 8.43GB** |
|
|
| ## Provenance |
|
|
| - Source: [`mistralai/Ministral-3-8B-Base-2512`](https://huggingface.co/mistralai/Ministral-3-8B-Base-2512) (BF16) |
| - Language model layers: **6-bit** affine quantization, group_size=64 |
| - Vision tower + multimodal projector: kept at full precision (not quantized) |
| - Blended average: **7.529 bits per weight** across all parameters |
| |
| ## Ministral 3 family |
| |
| | Model | Type | mlx-community (4-bit) | |
| |---|---|---| |
| | Ministral 3 3B Base 2512 | Base pre-trained | [mlx-community/Ministral-3-3B-Base-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-3B-Base-2512-4bit) | |
| | Ministral 3 3B Instruct 2512 | Instruct post-trained | [mlx-community/Ministral-3-3B-Instruct-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-3B-Instruct-2512-4bit) | |
| | Ministral 3 3B Reasoning 2512 | Reasoning capable | [mlx-community/Ministral-3-3B-Reasoning-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-3B-Reasoning-2512-4bit) | |
| | Ministral 3 8B Base 2512 | Base pre-trained | [mlx-community/Ministral-3-8B-Base-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-8B-Base-2512-4bit) | |
| | Ministral 3 8B Instruct 2512 | Instruct post-trained | [mlx-community/Ministral-3-8B-Instruct-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-8B-Instruct-2512-4bit) | |
| | Ministral 3 8B Reasoning 2512 | Reasoning capable | [mlx-community/Ministral-3-8B-Reasoning-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-8B-Reasoning-2512-4bit) | |
| | Ministral 3 14B Base 2512 | Base pre-trained | [mlx-community/Ministral-3-14B-Base-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-14B-Base-2512-4bit) | |
| | Ministral 3 14B Instruct 2512 | Instruct post-trained | [mlx-community/Ministral-3-14B-Instruct-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-14B-Instruct-2512-4bit) | |
| | Ministral 3 14B Reasoning 2512 | Reasoning capable | [mlx-community/Ministral-3-14B-Reasoning-2512-4bit](https://huggingface.co/mlx-community/Ministral-3-14B-Reasoning-2512-4bit) | |
| |
| ## Use with mlx |
| |
| ```bash |
| pip install -U mlx-vlm |
| ``` |
| |
| ```bash |
| python -m mlx_vlm.generate --model mlx-community/Ministral-3-8B-Base-2512-6bit --max-tokens 100 --temperature 0.0 --prompt "Describe this image." --image <path_to_image> |
| ``` |
| |
| For text-only prompts, omit `--image`. |
| |
| ## License |
| |
| Apache 2.0 — see the [original model card](https://huggingface.co/mistralai/Ministral-3-8B-Base-2512) for |
| the full license text and any usage terms. |
| |
