{"slug": "ibm-granite", "name": "IBM Granite", "category": "ai-models", "category_name": "AI models", "description": "", "recommenders": 12, "commenters": 17, "warners": 1, "score": 10.8, "rising": 0.05, "first": "2025-11-01", "latest": "2026-10-29", "why": "granite4 for speed and lower complexity (around the same as gpt20b).", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 14, "neutral": 5, "compare": 1, "warn": 1}, "spark": [2, 2, 0, 0, 1, 5, 0, 1, 0, 0, 1, 2], "listed": true, "loved": [], "caveat": "", "short_quote": "granite4 for speed and lower complexity (around the same as gpt20b).", "kind": "", "icon": true, "domain": "ibm.com", "github_repo": null, "official_url": "https://www.ibm.com/granite", "pricing": "", "aliases": ["Granite 4.1", "IBM Granite", "Granite", "Granite 4"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 804735200, "item_id": 50295950, "url": "https://news.ycombinator.com/item?id=50295950", "by": "salad-tycoon", "date": "2026-10-29", "version": null, "stance": "recommend", "quote": "granite is the model my m2 MacBook will be using to run this survival LLM", "excerpt": "Also granite is the model my m2 MacBook will be using to run this survival LLM I found on HN: https://www.brangus.io/survival-llm", "context": "Runs local LLMs on an M2 MacBook", "use_case": "Local LLM on Mac", "switched_from": "", "story": "We're releasing the weights of Mistral Large 4"}, {"id": 804291776, "item_id": 50268236, "url": "https://news.ycombinator.com/item?id=50268236", "by": "fultonn", "date": "2026-10-28", "version": null, "stance": "recommend", "quote": "I use granite and qwen together with a manually curated set of books, articles, etc. to get work done while disconnected.", "excerpt": "Cool project. I have some property that doesn't really get cell reception in the summer, and even in the winter the reception is at-best spotty. I use granite and qwen together with a manually curated set of books, articles, etc. to get work done while disconnected. I tend to update the set of resources before leaving town because I'm generally intentionally disconnecting to focus on deep work. Building a workflow th", "context": "works offline at a remote property", "use_case": "offline work with local models", "switched_from": "", "story": "Survival LLM: AI assistant and emergency library that runs on your own computer"}, {"id": 793097520, "item_id": 49568595, "url": "https://news.ycombinator.com/item?id=49568595", "by": "_zoltan_", "date": "2026-09-04", "version": null, "stance": "recommend", "quote": "Never heard of Granite?", "excerpt": "Never heard of Granite? https://huggingface.co/ibm-granite (I'm also an IBMer.)", "context": "IBMer", "use_case": "", "switched_from": "", "story": "IBM Bob"}, {"id": 777055424, "item_id": 48565964, "url": "https://news.ycombinator.com/item?id=48565964", "by": "EagnaIonat", "date": "2026-06-16", "version": "4.1", "stance": "recommend", "quote": "agentic granite4.1:3b works fine.", "excerpt": "> thers, like coding or more advanced agentic workflows can demand much more powerful models. You can do coding and agentic fine. For coding I use qwen3.6:35b-mlx and agentic granite4.1:3b works fine. These are the models I use. - granite4.1:3b - granite4.1:30b - gpt-oss:20b - gpt-oss:120b (less so now) - mistral-small3.2 - qwen3.6:35b-mlx There will always be use cases that don't sit on your laptop, but most of what", "context": "", "use_case": "local agentic tasks", "switched_from": "", "story": null}, {"id": 767480224, "item_id": 47967514, "url": "https://news.ycombinator.com/item?id=47967514", "by": "EagnaIonat", "date": "2026-04-30", "version": null, "stance": "recommend", "quote": "The weights are also open sourced as well.", "excerpt": "You could have found in 5 seconds. The weights are also open sourced as well. https://github.com/ibm-granite", "context": "", "use_case": "", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 767404496, "item_id": 47962781, "url": "https://news.ycombinator.com/item?id=47962781", "by": "whalesalad", "date": "2026-04-30", "version": "4.1", "stance": "warn", "quote": "it's just a bunch of other original content that has been chewed up and regurgitated into something \"new\"", "excerpt": "The thing is it's just a bunch of other original content that has been chewed up and regurgitated into something \"new\". Just show us the original content instead. This is by definition, slop. https://huggingface.co/blog/ibm-granite/granite-4-1", "context": "", "use_case": "", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}]}], "who": [{"by": "salad-tycoon", "context": "Runs local LLMs on an M2 MacBook", "url": "https://news.ycombinator.com/item?id=50295950", "date": "2026-10-29"}, {"by": "fultonn", "context": "works offline at a remote property", "url": "https://news.ycombinator.com/item?id=50268236", "date": "2026-10-28"}, {"by": "_zoltan_", "context": "IBMer", "url": "https://news.ycombinator.com/item?id=49568595", "date": "2026-09-04"}, {"by": "EagnaIonat", "context": "M4 Max 128GB Mac user", "url": "https://news.ycombinator.com/item?id=45820894", "date": "2025-11-05"}], "tags": [], "use_cases": ["local llm on mac", "offline work with local models", "local agentic tasks", "testing quantized model outputs", "local model trial", "open-weight ai model"], "compared": [{"slug": "gpt-oss", "name": "gpt-oss", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "qwen", "name": "Qwen", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "gemma", "name": "Gemma", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "nemotron", "name": "Nemotron", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=47960507", "title": "Granite 4.1: IBM's 8B Model Matching 32B MoE", "mentions": 7}, {"url": "https://news.ycombinator.com/item?id=46124324", "title": "IBM CEO says there is 'no way' spending on AI data centers will pay off", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=45771870", "title": "Ask HN: Who uses open LLMs and coding assistants locally? Share setup and laptop", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=50289418", "title": "We're releasing the weights of Mistral Large 4", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50264853", "title": "Survival LLM: AI assistant and emergency library that runs on your own computer", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49563851", "title": "IBM Bob", "mentions": 1}], "sentiment_by_month": {"2025-11": {"recommend": 2, "neutral": 1}, "2025-12": {"recommend": 2, "compare": 1}, "2026-01": {}, "2026-02": {}, "2026-03": {"recommend": 1}, "2026-04": {"recommend": 5, "warn": 1, "neutral": 1}, "2026-05": {"neutral": 1}, "2026-06": {"neutral": 2, "recommend": 1}, "2026-07": {}, "2026-08": {}, "2026-09": {"recommend": 1}, "2026-10": {"recommend": 2}}, "versions": [], "version_unnamed": 11, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 804735200, "item_id": 50295950, "url": "https://news.ycombinator.com/item?id=50295950", "by": "salad-tycoon", "date": "2026-10-29", "version": null, "stance": "recommend", "quote": "granite is the model my m2 MacBook will be using to run this survival LLM", "excerpt": "Also granite is the model my m2 MacBook will be using to run this survival LLM I found on HN: https://www.brangus.io/survival-llm", "context": "Runs local LLMs on an M2 MacBook", "use_case": "Local LLM on Mac", "switched_from": "", "story": "We're releasing the weights of Mistral Large 4"}, {"id": 804291776, "item_id": 50268236, "url": "https://news.ycombinator.com/item?id=50268236", "by": "fultonn", "date": "2026-10-28", "version": null, "stance": "recommend", "quote": "I use granite and qwen together with a manually curated set of books, articles, etc. to get work done while disconnected.", "excerpt": "Cool project. I have some property that doesn't really get cell reception in the summer, and even in the winter the reception is at-best spotty. I use granite and qwen together with a manually curated set of books, articles, etc. to get work done while disconnected. I tend to update the set of resources before leaving town because I'm generally intentionally disconnecting to focus on deep work. Building a workflow th", "context": "works offline at a remote property", "use_case": "offline work with local models", "switched_from": "", "story": "Survival LLM: AI assistant and emergency library that runs on your own computer"}, {"id": 793097520, "item_id": 49568595, "url": "https://news.ycombinator.com/item?id=49568595", "by": "_zoltan_", "date": "2026-09-04", "version": null, "stance": "recommend", "quote": "Never heard of Granite?", "excerpt": "Never heard of Granite? https://huggingface.co/ibm-granite (I'm also an IBMer.)", "context": "IBMer", "use_case": "", "switched_from": "", "story": "IBM Bob"}, {"id": 777055424, "item_id": 48565964, "url": "https://news.ycombinator.com/item?id=48565964", "by": "EagnaIonat", "date": "2026-06-16", "version": "4.1", "stance": "recommend", "quote": "agentic granite4.1:3b works fine.", "excerpt": "> thers, like coding or more advanced agentic workflows can demand much more powerful models. You can do coding and agentic fine. For coding I use qwen3.6:35b-mlx and agentic granite4.1:3b works fine. These are the models I use. - granite4.1:3b - granite4.1:30b - gpt-oss:20b - gpt-oss:120b (less so now) - mistral-small3.2 - qwen3.6:35b-mlx There will always be use cases that don't sit on your laptop, but most of what", "context": "", "use_case": "local agentic tasks", "switched_from": "", "story": null}, {"id": 776374675, "item_id": 48523417, "url": "https://news.ycombinator.com/item?id=48523417", "by": "lambda", "date": "2026-06-13", "version": null, "stance": "neutral", "quote": "Granite has much of the training pipeline and data open", "excerpt": "Nemotron is mostly open data. They only release a portions of their pre-training data. From https://docs.nvidia.com/nemotron/latest/nemotron/super3/pretrain.html Open-source data coverage: The released datasets cover an estimated 8\u201310T tokens (~40\u201350% of the internal 25T blend). Missing categories include code (~14% of blend), nemotron-cc-code (~2%), crawl++ (~2%), and academic text (~2%). Users should supplement wit", "context": "", "use_case": "", "switched_from": "", "story": "GLM 5.2 Is Out"}, {"id": 774007651, "item_id": 48375478, "url": "https://news.ycombinator.com/item?id=48375478", "by": "philipkglass", "date": "2026-06-02", "version": null, "stance": "neutral", "quote": "IBM publishes Granite.", "excerpt": "They aren't completely absent. Google keeps releasing Gemma models. Nvidia publishes Nemotron. Microsoft has their Phi series. IBM publishes Granite. Even OpenAI released a new open model (gpt-oss) less than a year ago. https://deepmind.google/models/gemma/gemma-4/ https://developer.nvidia.com/ai-models#:~:text=NVIDIA%20Nemotron,-The%20NVIDIA%20Nemotron https://www.microsoft.com/en-us/research/blog/phi-4-reasoning-vi", "context": "", "use_case": "", "switched_from": "", "story": "Trump signs executive order granting oversight of AI models"}, {"id": 767755568, "item_id": 47984723, "url": "https://news.ycombinator.com/item?id=47984723", "by": "EagnaIonat", "date": "2026-05-02", "version": null, "stance": "neutral", "quote": "https://huggingface.co/ibm-granite", "excerpt": "https://github.com/ibm-granite https://huggingface.co/ibm-granite I think if you were genuinely interested, you could have found this yourself.", "context": "", "use_case": "", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767480224, "item_id": 47967514, "url": "https://news.ycombinator.com/item?id=47967514", "by": "EagnaIonat", "date": "2026-04-30", "version": null, "stance": "recommend", "quote": "The weights are also open sourced as well.", "excerpt": "You could have found in 5 seconds. The weights are also open sourced as well. https://github.com/ibm-granite", "context": "", "use_case": "", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767475808, "item_id": 47967238, "url": "https://news.ycombinator.com/item?id=47967238", "by": "simonw", "date": "2026-04-30", "version": "4.1", "stance": "neutral", "quote": "This inspired me to give that a go", "excerpt": "This inspired me to give that a go: https://simonw.github.io/granite-4.1-3b-gguf-pelicans/", "context": "", "use_case": "Testing quantized model outputs", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767448209, "item_id": 47965513, "url": "https://news.ycombinator.com/item?id=47965513", "by": "2ndorderthought", "date": "2026-04-30", "version": "4.1", "stance": "recommend", "quote": "Qwen 3.6 35b a3b, granite4.1 8b, llama 3.2 3b.", "excerpt": "The biggest hint I have is set a budget. Then try some models out on either cloud instances or a computer you own. See if they work for you. Spec your machine accordingly. Some models I recommend trying to get a feel for what's out there. Qwen 3.6 35b a3b, granite4.1 8b, llama 3.2 3b. There are plenty of others but those give a good taste for different sizes and what they can do. If it's not enough then you are out m", "context": "", "use_case": "local model trial", "switched_from": "", "story": "Claude Code refuses requests or charges extra if your commits mention \"OpenClaw\""}, {"id": 767444496, "item_id": 47965281, "url": "https://news.ycombinator.com/item?id=47965281", "by": "EagnaIonat", "date": "2026-04-30", "version": "4.1", "stance": "recommend", "quote": "It's one of the few models on the market where the creator indemnifies it against copyright claims.", "excerpt": "Can you clarify what you mean? If you check HF you will see its Apache2 and the datasets were also permissive. It's one of the few models on the market where the creator indemnifies it against copyright claims. https://research.ibm.com/blog/granite-ethical-ai", "context": "", "use_case": "open-weight AI model", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767404496, "item_id": 47962781, "url": "https://news.ycombinator.com/item?id=47962781", "by": "whalesalad", "date": "2026-04-30", "version": "4.1", "stance": "warn", "quote": "it's just a bunch of other original content that has been chewed up and regurgitated into something \"new\"", "excerpt": "The thing is it's just a bunch of other original content that has been chewed up and regurgitated into something \"new\". Just show us the original content instead. This is by definition, slop. https://huggingface.co/blog/ibm-granite/granite-4-1", "context": "", "use_case": "", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767401904, "item_id": 47962619, "url": "https://news.ycombinator.com/item?id=47962619", "by": "woadwarrior01", "date": "2026-04-30", "version": "4.1", "stance": "recommend", "quote": "This makes then very token efficient and particularly well suited for local inference", "excerpt": "The most salient thing about these models is that they're non-reasoning models. This makes then very token efficient and particularly well suited for local inference where decoding is usually slower than with datacenter GPUs. Link to HF collection: https://huggingface.co/collections/ibm-granite/granite-41-language-models", "context": "", "use_case": "Local inference", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 767385808, "item_id": 47961613, "url": "https://news.ycombinator.com/item?id=47961613", "by": "pjmalandrino", "date": "2026-04-30", "version": "4.1", "stance": "recommend", "quote": "Very impressive series of SLM by IBM here.", "excerpt": "Very impressive series of SLM by IBM here. I have been using it with their Chunkless RAG concept and it is fitting very well! (for curious https://github.com/scub-france/Docling-Studio) I convinced that SLM are a real parto of solution for true integrated AI in process...", "context": "", "use_case": "chunkless RAG", "switched_from": "", "story": "Granite 4.1: IBM's 8B Model Matching 32B MoE"}, {"id": 764636881, "item_id": 47789805, "url": "https://news.ycombinator.com/item?id=47789805", "by": "Maxious", "date": "2026-04-16", "version": "4", "stance": "recommend", "quote": "like IBM did for granite 4", "excerpt": "Model authors are welcome to add support to llama.cpp before release like IBM did for granite 4 https://github.com/ggml-org/llama.cpp/pull/13550", "context": "", "use_case": "", "switched_from": "", "story": "Stop Using Ollama"}, {"id": 758166129, "item_id": 47385383, "url": "https://news.ycombinator.com/item?id=47385383", "by": "Western0", "date": "2026-03-15", "version": null, "stance": "recommend", "quote": "please look at old gemma 3n 4eb, and small IBM granite", "excerpt": "I only put this https://unsloth.ai/docs/models/qwen3.5 (please look at old gemma 3n 4eb, and small IBM granite)", "context": "", "use_case": "", "switched_from": "", "story": "Can I Run AI locally?"}, {"id": 739236704, "item_id": 46202294, "url": "https://news.ycombinator.com/item?id=46202294", "by": "cess11", "date": "2025-12-08", "version": null, "stance": "recommend", "quote": "I'd wager most people around here would have more use for Granite, however.", "excerpt": "They have? https://www.ibm.com/watson I'd wager most people around here would have more use for Granite, however. https://www.ibm.com/granite", "context": "", "use_case": "", "switched_from": "", "story": "IBM to acquire Confluent"}, {"id": 738226306, "item_id": 46139144, "url": "https://news.ycombinator.com/item?id=46139144", "by": "KronisLV", "date": "2025-12-03", "version": null, "stance": "compare", "quote": "It wasn't very promising when it came to benchmarks though", "excerpt": "You could have both worlds: an LLM model by IBM https://huggingface.co/ibm-granite/granite-4.0-h-small It wasn't very promising when it came to benchmarks though, go figure: https://artificialanalysis.ai/leaderboards/models", "context": "", "use_case": "Open-weight LLM", "switched_from": "", "story": "IBM CEO says there is 'no way' spending on AI data centers will pay off"}, {"id": 738198224, "item_id": 46137389, "url": "https://news.ycombinator.com/item?id=46137389", "by": "oedemis", "date": "2025-12-03", "version": null, "stance": "recommend", "quote": "here is very interesting research going https://huggingface.co/ibm-granite", "excerpt": "https://huggingface.co/ibm here is very interesting research going https://huggingface.co/ibm-granite", "context": "", "use_case": "", "switched_from": "", "story": "IBM CEO says there is 'no way' spending on AI data centers will pay off"}, {"id": 733134306, "item_id": 45820894, "url": "https://news.ycombinator.com/item?id=45820894", "by": "EagnaIonat", "date": "2025-11-05", "version": "4", "stance": "recommend", "quote": "granite4 for speed and lower complexity (around the same as gpt20b).", "excerpt": "The more recent LLMs work fine on an M1 mac. Can't speak for Windows/Linux. There was even a recent release of Granite4 that runs on a Raspberry Pi. https://github.com/Jewelzufo/granitepi-4-nano For my local work I use Ollama. (M4 Max 128GB) - gpt-oss. 20b or 120b depending on complexity of use cases. - granite4 for speed and lower complexity (around the same as gpt20b).", "context": "M4 Max 128GB Mac user", "use_case": "fast local LLM tasks", "switched_from": "", "story": "Developers are choosing older AI models"}, {"id": 732815873, "item_id": 45800992, "url": "https://news.ycombinator.com/item?id=45800992", "by": "fultonn", "date": "2025-11-03", "version": "4", "stance": "neutral", "quote": "I was referring to the smaller ones -- `granite4:micro`, `granite4-latest`, `granite4:350m`.", "excerpt": "> Edit: Hey which of the models on that page were you referring to? I was referring to the smaller ones -- `granite4:micro`, `granite4-latest`, `granite4:350m`. > I'm grabbing one now that's apparently double digit GB? You are probably downloading one of these two ids: `granite4:small-h` or `granite4:32b-a9b-h`. The \"small\" model _is_ small in relative terms, but is also the largest of the currently released granite ", "context": "", "use_case": "small local models", "switched_from": "", "story": "Ask HN: Who uses open LLMs and coding assistants locally? Share setup and laptop"}, {"id": 732522610, "item_id": 45782663, "url": "https://news.ycombinator.com/item?id=45782663", "by": "nickthegreek", "date": "2025-11-01", "version": null, "stance": "recommend", "quote": "small ibm granite worked surpassingly well", "excerpt": "just setup searxng yesterday and mcp for it in lm studio to be able to search the net for answers to simple queries. small ibm granite worked surpassingly well, while oss20b seemed to be looping searches.", "context": "", "use_case": "web search queries", "switched_from": "", "story": "Ask HN: Who uses open LLMs and coding assistants locally? Share setup and laptop"}]}