Ministral 3 8B
Ministral 3 8B is an open-source text generation model from Mistral with a 256,000-token context window.
MistralCompact open-source text generation from Mistral
Ministral 3 8B is an 8-billion-parameter language model developed by Mistral and released in December 2025. It is designed for text generation tasks and is available as an open-source model, meaning weights can be accessed and deployed independently. The model supports a 256,000-token context window, which allows it to process and reason over long documents within a single session, and it produces responses of up to 16,000 tokens.
Because of its relatively compact parameter count, Ministral 3 8B is suited for use cases where efficiency and deployability matter, such as summarization, question answering, instruction following, and conversational applications. Its open-source status makes it accessible for developers who want to self-host or fine-tune the model for specific tasks. It is served first-party through Mistral and is available directly on MindStudio without requiring separate API credentials.
What Ministral 3 8B supports
Long Context Processing
Handles up to 256,000 tokens of input in a single session, enabling analysis of lengthy documents, codebases, or multi-turn conversations without truncation.
Text Generation
Generates coherent, instruction-following text responses across tasks such as summarization, Q&A, and drafting, with a maximum response size of 16,000 tokens.
Open-Source Weights
Released as an open-source model, allowing developers to download, self-host, or fine-tune the 8B-parameter weights for custom deployments.
Instruction Following
Trained to follow natural language instructions, making it suitable for chat interfaces, automated pipelines, and agent-style task execution.
Efficient Inference
At 8 billion parameters, the model is designed to run with lower compute requirements than larger variants, supporting faster response times in resource-constrained environments.
Ready to build with Ministral 3 8B?
Get Started FreeBenchmark scores
Scores represent accuracy — the percentage of questions answered correctly on each test.
| Benchmark | What it tests | Score |
|---|---|---|
| MMLU-Pro | Expert knowledge across 14 academic disciplines | 64.2% |
| GPQA Diamond | PhD-level science questions (biology, physics, chemistry) | 47.1% |
| LiveCodeBench | Real-world coding tasks from recent competitions | 30.3% |
| HLE | Questions that challenge frontier models across many domains | 4.3% |
| SciCode | Scientific research coding and numerical methods | 20.8% |
Common questions about Ministral 3 8B
What is the context window size for Ministral 3 8B?
Ministral 3 8B supports a context window of 256,000 tokens, allowing it to process very long documents or extended conversations in a single session.
What is the maximum response length?
The model can generate responses of up to 16,000 tokens per request.
Is Ministral 3 8B open source?
Yes, Ministral 3 8B is tagged as an open-source model, meaning its weights are publicly available for download, self-hosting, and fine-tuning.
Who publishes Ministral 3 8B and how is it provided on MindStudio?
The model is published by Mistral and is provided first-party on MindStudio, so no separate API key or external account is required to use it.
What types of tasks is Ministral 3 8B best suited for?
The model is designed for general text generation tasks including summarization, question answering, instruction following, and conversational applications, particularly in settings where a compact, deployable model is preferred.
What people think about Ministral 3 8B
Community discussion around Ministral 3 8B on Reddit has been generally positive, with users in the LocalLLaMA subreddit noting its release as part of a broader wave of Mistral model launches. The release thread received 280 upvotes and 61 comments, reflecting meaningful interest from the local inference community.
Some discussion touched on Mistral's rapid model release cadence rather than focusing on specific benchmarks or limitations of the 3 8B variant. Users interested in base models and local deployment appear to be the primary audience engaging with this model.
Mistral AI drops 3x as many LLMs in a single week as OpenAI did in 6 years
Ministral-3 has been released
Google Gemini 3.1 Pro Preview Soon?
Looking for a Base Model
Parameters & options
Explore similar models
MistralMinistral 3 3B
Ministral 3 3B is a 3-billion-parameter open-source text generation model from Mistral with a 256,000-token context window.
MistralMinistral 3 14B
Ministral 3 14B is an open-source text generation model from Mistral with a 256,000-token context window.
MistralMistral Large 3
Mistral Large 3 is an open-source text generation model from Mistral with a 256,000 token context window.
IBMGranite 4.2 30B
TencentHy3
Hy3 is a text generation model from Tencent with a 262,144-token context window, available via DeepInfra.
XiaomiMiMo V2.5
MistralMistral Codestral
Mistral Codestral is a code-focused text generation model from Mistral with a 32,000 token context window.
MistralMistral 8x7b
Mistral 8x7b is a text generation model from Mistral AI using a mixture-of-experts architecture with a 32,768 token context window.
Start building with Ministral 3 8B
No API keys required. Create AI-powered workflows with Ministral 3 8B in minutes — free.