Get the latest setups
Models/Ministral 3 14B Reasoning 2512
Text / mistralai

Ministral 3 14B Reasoning 2512

Ministral 3 14B Reasoning 2512 is the reasoning post-trained version of the largest model in the Ministral 3 family, released by mistralai under the apache-2.0 license. It combines a 13.5B language model with a 0.4B vision encoder, supports vision, multilingual input across eleven listed languages, system prompts, native function calling and JSON outputting, and a 256k context window. The model is described as edge-optimized and is said to fit in 32GB of VRAM in BF16, or less than 24GB of RAM/VRAM when quantized. It is intended for reasoning-heavy tasks such as math, coding, and STEM, as well as private or local deployments, agentic uses, and fine-tuning. The model card recommends using the provided system prompt, keeping reasoning traces in context for multi-turn interactions, a temperature of 1 for most environments, limiting the number of tools, and maintaining an aspect ratio close to 1:1 for images.

EstimatedRepository opened Oct 31, 2025Source checked 9/25/2026Version: 51f9210f
LOCALRENTED GPUOPEN WEIGHTS
Freshness 23 out of 100. Repository age: 379 days old, Oct 31, 2025. Measures age, not quality.
Popularity 72 out of 100. 359,619 downloads in 30 days; 176 lifetime likes. Checked Nov 13, 2026. Measures adoption, not quality.
What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source
Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source
Scores as of Nov 14, 2026. Statistics checked Nov 13, 2026. Statistics are marked for refresh after 72 hours.Full scoring method
WORKFLOWS FOR THIS MODEL 1

What do you want to do?

Choose a task. Get the software, files, settings, and steps to run it.

MAKE IT PERSONALChoose your hardware
AI agent

Build a local AI agent with Ministral 3 Reasoning

Run an official Ministral 3 Reasoning GGUF on your computer, allow one bounded local lookup, and inspect the tool result and final answer.

Freshness 23 out of 100. Repository age: 379 days old, Oct 31, 2025. Measures age, not quality.
Popularity 72 out of 100. 359,619 downloads in 30 days; 176 lifetime likes. Checked Nov 13, 2026. Measures adoption, not quality.
OllamaPython
3 packagesmacOS / Windows / Linux
Open agent guide

Open a guide for its software, hardware requirements, and setup steps.

At a glance

Parameters
13.95B
Architecture
mistral3
License
apache-2.0
Disk space
26 GB
Software
vLLM
View the model source
MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

  • Private or custom chat and AI assistant deployments in constrained environments
  • Advanced local agentic use cases with native function calling and JSON outputting
  • Math, coding, and STEM-related reasoning tasks
  • Fine-tuning and specialization on domain-specific data
  • Vision-enabled analysis of images alongside text

Tradeoffs

  • The source does not specify hardware requirements beyond fitting in 32GB of VRAM in BF16 and under 24GB of RAM/VRAM when quantized.
  • No performance benchmarks, testing results, or compatibility details are provided.
  • The model card advises against overly thin or wide images and recommends cropping to an aspect ratio near 1:1 for vision.
  • The model card recommends limiting the number of tools to avoid overloading the model in agentic use cases.
  • No information is provided about supported inference frameworks beyond the listed library_name: vllm and inference: false.

Ways to run it

vLLM · Linux

Repository-specific command in publisher documentation; confirm dependencies and hardware in the source.

vllm serve mistralai/Ministral-3-14B-Reasoning-2512 --tensor-parallel-size 2 --tokenizer_mode mistral --config_format mistral --load_format mistral --enable-auto-tool-choice --tool-call-parser mistral --reasoning-parser mistral
Official instructions

Explore its uses

SAFETENSORS

Keep exploring

Text
Prism ML

Bonsai 2 · 27B

A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.

Freshness 80 out of 100. Release age: 58 days old, Sep 17, 2026. Measures age, not quality.
Popularity 85 out of 100. 1,522,102 downloads in 30 days; 3,006 lifetime likes. Checked Nov 13, 2026. Measures adoption, not quality.
MODEL SIZE27.36B
LOCALRENTED GPU
Released Sep 17, 2026
Text
DeepSeek

DeepSeek V4.1 Flash

An open-weight model for reasoning, coding, agents, and image understanding. DeepSeek V4.1 Flash combines a 552B backbone with sparse activation and up to one million tokens of context. Choose the original checkpoint or a community uncensored variant below.

Freshness 78 out of 100. Repository age: 65 days old, Sep 10, 2026. Measures age, not quality.
Popularity 86 out of 100. 1,772,553 downloads in 30 days; 4,959 lifetime likes. Checked Nov 13, 2026. Measures adoption, not quality.
MODEL SIZE552B backbone · 8B active prefill / 16B active decode
LOCALRENTED GPU
Repository opened Sep 10, 2026
Text
Qwen

Qwen3.8

Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats
LOCALRENTED GPU
Compare versions →