{"slug": "modernbert", "name": "ModernBERT", "category": "ai-models", "category_name": "AI models", "description": "", "recommenders": 3, "commenters": 6, "warners": 0, "score": 3.556, "rising": -0.98, "first": "2026-03-03", "latest": "2026-10-11", "why": "ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 3, "compare": 0, "warn": 0}, "spark": [0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin", "kind": "", "icon": true, "domain": null, "github_repo": null, "official_url": "https://github.com/AnswerDotAI/ModernBERT", "pricing": "", "aliases": ["ModernBERT"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 779666848, "item_id": 48729178, "url": "https://news.ycombinator.com/item?id=48729178", "by": "kamranjon", "date": "2026-06-29", "version": null, "stance": "recommend", "quote": "ModernBERT is amazing if you haven\u2019t used it before, worth spending time with", "excerpt": "If anyone is interested in doing something seriously useful with these neural cores, there is this incredible write up on getting ModernBERT running on them: https://stephenpanaro.com/blog/modernbert-on-apple-neural-engine Really wish this author would blog more, this piece is incredible and includes the code. Also ModernBERT is amazing if you haven\u2019t used it before, worth spending time with - have used it myself for", "context": "", "use_case": "text classification", "switched_from": "", "story": "Apple Neural Engine: Architecture, Programming, and Performance"}, {"id": 778026128, "item_id": 48626633, "url": "https://news.ycombinator.com/item?id=48626633", "by": "kamranjon", "date": "2026-06-21", "version": null, "stance": "recommend", "quote": "ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin", "excerpt": "I have! I recently compared Gemma 1b to ModernBERT Large for a binary classification task and ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin by the end of training. It seems the bidirectional encoder only architecture works really well for classification tasks, and I think it is related to being bidirectional whereas decoder only models like Gemma (or Qwen) ca", "context": "", "use_case": "binary classification", "switched_from": "", "story": "Good results fine tuning a local LLM like Qwen 3:0.6B to categorize questions"}, {"id": 768292672, "item_id": 48018292, "url": "https://news.ycombinator.com/item?id=48018292", "by": "jayshah5696", "date": "2026-05-04", "version": null, "stance": "recommend", "quote": "embeddings are easy to fine tune. Try modern bert.", "excerpt": "embeddings are easy to fine tune. Try modern bert.", "context": "", "use_case": "fine-tuning embeddings", "switched_from": "", "story": "Ask HN: Best Embedding Models?"}, {"id": 755784196, "item_id": 47236512, "url": "https://news.ycombinator.com/item?id=47236512", "by": "Horos", "date": "2026-03-03", "version": null, "stance": "recommend", "quote": "small encoder-only classifiers (DeBERTa-small, ModernBERT)", "excerpt": "Interesting gap between surface scanning (6.6%) and AI deep audit (16.4%). Two concerns with the AI audit approach. First, the defense LLM is itself an attack surface \u2014 we're already seeing payloads crafted specifically to bypass LLM-based guardrails. If the guardian is injectable, you've added a vulnerability to your security stack. Second, the Mindgard paper from late 2025 tested 12 character injection techniques a", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: What are you working on? (March 2026)"}]}], "caveats": [], "who": [{"by": "earino", "context": "Teaches a class on fine-tuning models", "url": "https://news.ycombinator.com/item?id=49890557", "date": "2026-09-29"}], "tags": [], "use_cases": ["text classifier base model", "fine-tuning course examples", "classifying ai-related articles", "text classification", "binary classification", "fine-tuning embeddings"], "compared": [], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=50037949", "title": "Build your own decision model", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49883844", "title": "Jeff \u2013 Jev-compatible 0.8B decision models, trained at home, ~30 ms", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49659647", "title": "Show HN: Hacker News, without AI", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48702825", "title": "Apple Neural Engine: Architecture, Programming, and Performance", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48623434", "title": "Good results fine tuning a local LLM like Qwen 3:0.6B to categorize questions", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48017668", "title": "Ask HN: Best Embedding Models?", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {"recommend": 1}, "2026-04": {}, "2026-05": {"recommend": 1}, "2026-06": {"recommend": 1}, "2026-07": {}, "2026-08": {}, "2026-09": {"neutral": 2}, "2026-10": {"neutral": 1}}, "versions": [], "version_unnamed": 6, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 800753744, "item_id": 50047109, "url": "https://news.ycombinator.com/item?id=50047109", "by": "make3", "date": "2026-10-11", "version": null, "stance": "neutral", "quote": "", "excerpt": "The real guide is here: https://huggingface.co/convaiinnovations/laya", "context": "", "use_case": "text classifier base model", "switched_from": "", "story": "Build your own decision model"}, {"id": 798248913, "item_id": 49890557, "url": "https://news.ycombinator.com/item?id=49890557", "by": "earino", "date": "2026-09-29", "version": null, "stance": "neutral", "quote": "using the same ModernBERT stuff", "excerpt": "Thanks for the vouching. I don't really understand what I did to anger the algorithms! But yeah, I think that that's the real sticking point. If the value prop of jev is \"plug it in anywhere and it generally does smart things\" you're absolutely right that Laya is Not A Thing. I've spent a bunch of time working on fine tuning models (i teach a class on it! https://earino.github.io/applied-deep-learning/ using the same", "context": "Teaches a class on fine-tuning models", "use_case": "Fine-tuning course examples", "switched_from": "", "story": "Jeff \u2013 Jev-compatible 0.8B decision models, trained at home, ~30 ms"}, {"id": 794564817, "item_id": 49660301, "url": "https://news.ycombinator.com/item?id=49660301", "by": "postalcoder", "date": "2026-09-11", "version": null, "stance": "neutral", "quote": "a Modern Bert-based classifier fine-tuned to detect AI-related content", "excerpt": "Just a few high level points about how this works. It works in multiple stages: 1. I apply a domain and keyword filter to the feed 2. The content of the remaining articles are run twice daily through a Modern Bert-based classifier fine-tuned to detect AI-related content (~8000 training examples)[0]. It also filters out Github repos that contain AI authorship. My backend scans: - Commit messages for agent attribution ", "context": "", "use_case": "classifying AI-related articles", "switched_from": "", "story": "Show HN: Hacker News, without AI"}, {"id": 779666848, "item_id": 48729178, "url": "https://news.ycombinator.com/item?id=48729178", "by": "kamranjon", "date": "2026-06-29", "version": null, "stance": "recommend", "quote": "ModernBERT is amazing if you haven\u2019t used it before, worth spending time with", "excerpt": "If anyone is interested in doing something seriously useful with these neural cores, there is this incredible write up on getting ModernBERT running on them: https://stephenpanaro.com/blog/modernbert-on-apple-neural-engine Really wish this author would blog more, this piece is incredible and includes the code. Also ModernBERT is amazing if you haven\u2019t used it before, worth spending time with - have used it myself for", "context": "", "use_case": "text classification", "switched_from": "", "story": "Apple Neural Engine: Architecture, Programming, and Performance"}, {"id": 778026128, "item_id": 48626633, "url": "https://news.ycombinator.com/item?id=48626633", "by": "kamranjon", "date": "2026-06-21", "version": null, "stance": "recommend", "quote": "ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin", "excerpt": "I have! I recently compared Gemma 1b to ModernBERT Large for a binary classification task and ModernBERT was the clear winner. It learned faster and performed the task better by a significant margin by the end of training. It seems the bidirectional encoder only architecture works really well for classification tasks, and I think it is related to being bidirectional whereas decoder only models like Gemma (or Qwen) ca", "context": "", "use_case": "binary classification", "switched_from": "", "story": "Good results fine tuning a local LLM like Qwen 3:0.6B to categorize questions"}, {"id": 768292672, "item_id": 48018292, "url": "https://news.ycombinator.com/item?id=48018292", "by": "jayshah5696", "date": "2026-05-04", "version": null, "stance": "recommend", "quote": "embeddings are easy to fine tune. Try modern bert.", "excerpt": "embeddings are easy to fine tune. Try modern bert.", "context": "", "use_case": "fine-tuning embeddings", "switched_from": "", "story": "Ask HN: Best Embedding Models?"}, {"id": 755784196, "item_id": 47236512, "url": "https://news.ycombinator.com/item?id=47236512", "by": "Horos", "date": "2026-03-03", "version": null, "stance": "recommend", "quote": "small encoder-only classifiers (DeBERTa-small, ModernBERT)", "excerpt": "Interesting gap between surface scanning (6.6%) and AI deep audit (16.4%). Two concerns with the AI audit approach. First, the defense LLM is itself an attack surface \u2014 we're already seeing payloads crafted specifically to bypass LLM-based guardrails. If the guardian is injectable, you've added a vulnerability to your security stack. Second, the Mindgard paper from late 2025 tested 12 character injection techniques a", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: What are you working on? (March 2026)"}]}