{"slug": "ministral", "name": "Ministral", "category": "ai-models", "category_name": "AI models", "description": "", "recommenders": 6, "commenters": 11, "warners": 0, "score": 6.395, "rising": -1.96, "first": "2025-12-02", "latest": "2026-05-22", "why": "I tested locally using Ministral 3B and it was promising.", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 6, "neutral": 6, "compare": 0, "warn": 0}, "spark": [0, 1, 0, 3, 1, 0, 1, 0, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "I tested locally using Ministral 3B and it was promising.", "kind": "", "icon": true, "family": "mistral", "domain": "mistral.ai", "github_repo": null, "official_url": "https://mistral.ai", "pricing": "", "aliases": ["Ministral 3B", "Ministral 14B", "Ministral 3 8B", "Ministral 3", "Ministral"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 771804944, "item_id": 48237809, "url": "https://news.ycombinator.com/item?id=48237809", "by": "sometimelurker", "date": "2026-05-22", "version": "3B", "stance": "recommend", "quote": "I run Ministral-3b on my hardware as my go-to", "excerpt": "sorry to nitpick (I totally agree with what ur saying btw, I run Ministral-3b on my hardware as my go-to bc I don't usually need the \"smartest and most expensive models\") > This is where open source models are important open-weights, the training data isn't public", "context": "", "use_case": "Local everyday model", "switched_from": "", "story": "The current AI pricing was always going to go away"}, {"id": 758504272, "item_id": 47406517, "url": "https://news.ycombinator.com/item?id=47406517", "by": "Adrig", "date": "2026-03-16", "version": null, "stance": "recommend", "quote": "it performed the best in my benchmarks", "excerpt": "I used Ministral for data cleaning. I was surprised: even tho it was the cheapest option (against other small models from Anthropic) it performed the best in my benchmarks.", "context": "", "use_case": "Data cleaning", "switched_from": "", "story": "Mistral Releases Leanstral"}, {"id": 755128560, "item_id": 47195535, "url": "https://news.ycombinator.com/item?id=47195535", "by": "lioeters", "date": "2026-02-28", "version": "3", "stance": "recommend", "quote": "fast enough that I never wait for it", "excerpt": "I've been running local language models on an existing laptop with 8GB GPU, currently using ministral-3:8b. It's faster than other models of similar size I used previously, fast enough that I never wait for it, rather have to scroll back to read the full output. So far I'm using it conversationally, and scripting with tools. I wrote a simple chat interface / REPL in the terminal. But it's not integrated with code edi", "context": "runs local models on 8GB GPU laptop", "use_case": "Local chat and code help", "switched_from": "", "story": "How do I cancel my ChatGPT subscription?"}, {"id": 753425265, "item_id": 47089079, "url": "https://news.ycombinator.com/item?id=47089079", "by": "brainless", "date": "2026-02-20", "version": "3B", "stance": "recommend", "quote": "I tested locally using Ministral 3B and it was promising.", "excerpt": "I know it is not easy to see the benefits of small models easily but this is what I am building for (1). I created a product for Google Gemini 3 Hackathon and I used Gemini 3 Flash (2). I tested locally using Ministral 3B and it was promising. Definitely will need work. But 8B/14B may give awesome results. I am building a data extraction software on top of emails, attachments, cloud/local files. I use a reverse templ", "context": "", "use_case": "local data extraction", "switched_from": "", "story": "The path to ubiquitous AI (17k tokens/sec)"}, {"id": 750912289, "item_id": 46932018, "url": "https://news.ycombinator.com/item?id=46932018", "by": "lodovic", "date": "2026-02-07", "version": null, "stance": "recommend", "quote": "I was pleasantly surprised by how fast and accurate these small models have become.", "excerpt": "I made something similar to this project, and tested it against a few 3B and 8B models (Qwen and Ministral, both the instruction and the reasoning variants). I was pleasantly surprised by how fast and accurate these small models have become. I can ask it things like \"check out this repo and build it\", and with a Ralph strategy eventually it will succeed, despite the small context size.", "context": "", "use_case": "Small local models", "switched_from": "", "story": "Show HN: LocalGPT \u2013 A local-first AI assistant in Rust with persistent memory"}]}], "caveats": [], "who": [{"by": "lioeters", "context": "runs local models on 8GB GPU laptop", "url": "https://news.ycombinator.com/item?id=47195535", "date": "2026-02-28"}], "tags": [], "use_cases": ["local llm", "local everyday model", "data cleaning", "local chat and code help", "local data extraction", "small local models"], "compared": [{"slug": "qwen", "name": "Qwen", "co": 3, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=48234391", "title": "The current AI pricing was always going to go away", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48143284", "title": "Cut Off: why access to frontier AI will soon be scarce and selective", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47404796", "title": "Mistral Releases Leanstral", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47190997", "title": "How do I cancel my ChatGPT subscription?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47086181", "title": "The path to ubiquitous AI (17k tokens/sec)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=46930391", "title": "Show HN: LocalGPT \u2013 A local-first AI assistant in Rust with persistent memory", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {"recommend": 1, "neutral": 1}, "2026-01": {"neutral": 4}, "2026-02": {"recommend": 3}, "2026-03": {"recommend": 1}, "2026-04": {}, "2026-05": {"neutral": 1, "recommend": 1}, "2026-06": {}, "2026-07": {}, "2026-08": {}, "2026-09": {}, "2026-10": {}}, "versions": [], "version_unnamed": 2, "family_ref": {"slug": "mistral", "name": "Mistral"}, "runs_on": null, "lines": [], "quotes": [{"id": 771804944, "item_id": 48237809, "url": "https://news.ycombinator.com/item?id=48237809", "by": "sometimelurker", "date": "2026-05-22", "version": "3B", "stance": "recommend", "quote": "I run Ministral-3b on my hardware as my go-to", "excerpt": "sorry to nitpick (I totally agree with what ur saying btw, I run Ministral-3b on my hardware as my go-to bc I don't usually need the \"smartest and most expensive models\") > This is where open source models are important open-weights, the training data isn't public", "context": "", "use_case": "Local everyday model", "switched_from": "", "story": "The current AI pricing was always going to go away"}, {"id": 770826439, "item_id": 48176652, "url": "https://news.ycombinator.com/item?id=48176652", "by": "rsolva", "date": "2026-05-18", "version": "14B", "stance": "neutral", "quote": "Mostly Ministral 14B", "excerpt": "Nowadays I use our local setup 95% of the time, but it is not that long since that flipped for me personally. Context: I have a $20 Claude Code subscription, and have used it for a handfull of small-ish projects the last year, in parallel with local models on my AMD 9700XTX (24GB) at home. Mostly Ministral 14B and more recently Qwen3.6 27B Dense 4q. Historically, the tooling (interferens engines and harness) has been", "context": "", "use_case": "", "switched_from": "", "story": "Cut Off: why access to frontier AI will soon be scarce and selective"}, {"id": 758504272, "item_id": 47406517, "url": "https://news.ycombinator.com/item?id=47406517", "by": "Adrig", "date": "2026-03-16", "version": null, "stance": "recommend", "quote": "it performed the best in my benchmarks", "excerpt": "I used Ministral for data cleaning. I was surprised: even tho it was the cheapest option (against other small models from Anthropic) it performed the best in my benchmarks.", "context": "", "use_case": "Data cleaning", "switched_from": "", "story": "Mistral Releases Leanstral"}, {"id": 755128560, "item_id": 47195535, "url": "https://news.ycombinator.com/item?id=47195535", "by": "lioeters", "date": "2026-02-28", "version": "3", "stance": "recommend", "quote": "fast enough that I never wait for it", "excerpt": "I've been running local language models on an existing laptop with 8GB GPU, currently using ministral-3:8b. It's faster than other models of similar size I used previously, fast enough that I never wait for it, rather have to scroll back to read the full output. So far I'm using it conversationally, and scripting with tools. I wrote a simple chat interface / REPL in the terminal. But it's not integrated with code edi", "context": "runs local models on 8GB GPU laptop", "use_case": "Local chat and code help", "switched_from": "", "story": "How do I cancel my ChatGPT subscription?"}, {"id": 753425265, "item_id": 47089079, "url": "https://news.ycombinator.com/item?id=47089079", "by": "brainless", "date": "2026-02-20", "version": "3B", "stance": "recommend", "quote": "I tested locally using Ministral 3B and it was promising.", "excerpt": "I know it is not easy to see the benefits of small models easily but this is what I am building for (1). I created a product for Google Gemini 3 Hackathon and I used Gemini 3 Flash (2). I tested locally using Ministral 3B and it was promising. Definitely will need work. But 8B/14B may give awesome results. I am building a data extraction software on top of emails, attachments, cloud/local files. I use a reverse templ", "context": "", "use_case": "local data extraction", "switched_from": "", "story": "The path to ubiquitous AI (17k tokens/sec)"}, {"id": 750912289, "item_id": 46932018, "url": "https://news.ycombinator.com/item?id=46932018", "by": "lodovic", "date": "2026-02-07", "version": null, "stance": "recommend", "quote": "I was pleasantly surprised by how fast and accurate these small models have become.", "excerpt": "I made something similar to this project, and tested it against a few 3B and 8B models (Qwen and Ministral, both the instruction and the reasoning variants). I was pleasantly surprised by how fast and accurate these small models have become. I can ask it things like \"check out this repo and build it\", and with a Ralph strategy eventually it will succeed, despite the small context size.", "context": "", "use_case": "Small local models", "switched_from": "", "story": "Show HN: LocalGPT \u2013 A local-first AI assistant in Rust with persistent memory"}, {"id": 747869426, "item_id": 46741839, "url": "https://news.ycombinator.com/item?id=46741839", "by": "andhuman", "date": "2026-01-23", "version": "14B", "stance": "neutral", "quote": "mistral ministral 14b as LLM", "excerpt": "I built this recently. I used nvidia parakeet as STT, open wake word as the wake word detection, mistral ministral 14b as LLM and pocket tts for tts. Fits snugly in my 16 gb VRAM. Pocket is small and fast and has good enough voice cloning. I first used the chatterbox turbo model, which perform better and even supported some simple paralinguistic word like (chuckle) that made it more fun, but it was just a bit too big", "context": "", "use_case": "local LLM", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 746910528, "item_id": 46681908, "url": "https://news.ycombinator.com/item?id=46681908", "by": "piyh", "date": "2026-01-19", "version": "3", "stance": "neutral", "quote": "https://docs.mistral.ai/models/ministral-3-8b-25-12", "excerpt": "https://docs.mistral.ai/models/ministral-3-8b-25-12", "context": "", "use_case": "", "switched_from": "", "story": "GLM-4.7-Flash"}, {"id": 745209376, "item_id": 46575586, "url": "https://news.ycombinator.com/item?id=46575586", "by": "lioeters", "date": "2026-01-11", "version": "3", "stance": "neutral", "quote": "A recent local model I tried is Ministral 3 from a month ago.", "excerpt": "A recent local model I tried is Ministral 3 from a month ago. https://mistral.ai/news/mistral-3 Vision: Enables the model to analyze images and provide insights based on visual content, in addition to text. Multilingual: Supports dozens of languages, including English, French, Spanish, German, Italian, Portuguese, Dutch, Chinese, Japanese, Korean, Arabic. ... Agentic: Offers best-in-class agentic capabilities with na", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: What happened to self-hosted models?"}, {"id": 744916658, "item_id": 46557291, "url": "https://news.ycombinator.com/item?id=46557291", "by": "armcat", "date": "2026-01-09", "version": "3", "stance": "neutral", "quote": "I swap between ministral-3 and Qwen3", "excerpt": "I have a 5070 in my rig. What I'm running is Kokoro in a Python/FastAPI backend - I also use local quantized models (I swap between ministral-3 and Qwen3) as \"the brains\" (offload to GPT-5.2 inc. web search for \"complex\" tasks or those requiring the web). In the backend I use Kokoro and generate wav bytes that I send to the frontend. The frontend is just a simple HTML page with a textbox and a button, invoking a `fet", "context": "", "use_case": "Local LLM", "switched_from": "", "story": "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"}, {"id": 738457040, "item_id": 46153565, "url": "https://news.ycombinator.com/item?id=46153565", "by": "troyvit", "date": "2025-12-04", "version": "3B", "stance": "neutral", "quote": "there's a Mistral 3B model that runs locally in a browser", "excerpt": "The only difference is that memory prices skyrocketing is a temporary thing resulting from a spike in demand from incompetent AI megalomaniacs like Sam Altman who don't know how to run a company and are desperate to scale because that's the only kind of sustainability they understand. Once the market either absorbs that demand (if it's real) or else over-produces for it, RAM prices are going to either slowly come bac", "context": "", "use_case": "", "switched_from": "", "story": "Everyone in Seattle hates AI"}, {"id": 737983616, "item_id": 46123976, "url": "https://news.ycombinator.com/item?id=46123976", "by": "simonw", "date": "2025-12-02", "version": "3B", "stance": "recommend", "quote": "There's a very cool demo of that here", "excerpt": "The 3B vision model runs in the browser (after a 3GB model download). There's a very cool demo of that here: https://huggingface.co/spaces/mistralai/Ministral_3B_WebGPU Pelicans are OK but not earth-shattering: https://simonwillison.net/2025/Dec/2/introducing-mistral-3/", "context": "", "use_case": "Vision model running in browser", "switched_from": "", "story": "Mistral 3 family of models released"}]}