{"slug": "mtplx", "name": "MTPLX", "category": "ai", "category_name": "AI models and assistants", "description": "", "recommenders": 3, "commenters": 4, "warners": 1, "score": 1.191, "rising": 1.67, "first": "2026-06-11", "latest": "2026-10-08", "why": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 1, "compare": 0, "warn": 0}, "spark": [0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "icon": false, "domain": null, "github_repo": null, "official_url": "", "pricing": "", "aliases": ["MTPLX"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 793811553, "item_id": 49613222, "url": "https://news.ycombinator.com/item?id=49613222", "by": "bellowsgulch", "date": "2026-09-08", "stance": "recommend", "quote": "I\u2019ll go use that now that I know.", "excerpt": "Thanks, man! I\u2019ll go use that now that I know. llama-server the last time I used it for inference with this model wasn\u2019t able to produce work fast enough to reach those numbers.", "context": "", "use_case": "faster local inference", "switched_from": "llama-server", "story": "Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses"}, {"id": 789224400, "item_id": 49326525, "url": "https://news.ycombinator.com/item?id=49326525", "by": "suoloordi", "date": "2026-08-16", "stance": "recommend", "quote": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "excerpt": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "context": "36GB MacBook Pro user", "use_case": "running Qwen locally on Mac", "switched_from": "", "story": "Qwen 3.8 27B is excellent, but it defaults to overthinking things"}, {"id": 775928928, "item_id": 48495558, "url": "https://news.ycombinator.com/item?id=48495558", "by": "rrevi", "date": "2026-06-11", "stance": "recommend", "quote": "I have been using MTPLX (replaced oMLX for me)", "excerpt": "I have been using MTPLX (replaced oMLX for me) on my M5 Pro with 64GB with Qwen3.6 27B Optimized Quality (getting about 20-25 generated tps) and along with OpenCode have been enjoying the experience very much :) HUGE shoutout to Youssof! Thank you.", "context": "M5 Pro 64GB user", "use_case": "local LLM inference", "switched_from": "oMLX", "story": "Mtplx 1.0.0"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 795681441, "item_id": 49730090, "url": "https://news.ycombinator.com/item?id=49730090", "by": "bellowsgulch", "date": "2026-09-16", "stance": "warn", "quote": "I still found the general MLX MTP acceleration to be so poor", "excerpt": "I so wanted to be able to use Qwen3.8-27B locally on my fully loaded M1 Max 64GB when spider-mario recommended I use MTPLX[1], but I still found the general MLX MTP acceleration to be so poor, I'd rather just use cheap tokens from OpenCode Zen and Go. It's just too slow of a model. I know, I know there's new hardware, but my business paid like 4-5 grand for this MBP at the time, and I just don't feel the need to pay ", "context": "", "use_case": "local model acceleration", "switched_from": "", "story": "Show HN: Swift-Qwen3.8-27B, -58.3% thinking, x1.95 speed, accuracy of xhigh"}]}], "who": [{"by": "suoloordi", "context": "36GB MacBook Pro user", "url": "https://news.ycombinator.com/item?id=49326525", "date": "2026-08-16"}, {"by": "rrevi", "context": "M5 Pro 64GB user", "url": "https://news.ycombinator.com/item?id=48495558", "date": "2026-06-11"}], "tags": [], "use_cases": ["running local models", "local model acceleration", "faster local inference", "running qwen locally on mac", "local llm inference"], "compared": [{"slug": "llama-server", "name": "llama-server", "co": 1, "switched_from": 1, "switched_to": 0}, {"slug": "omlx", "name": "oMLX", "co": 1, "switched_from": 1, "switched_to": 0}, {"slug": "opencode", "name": "OpenCode", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "qwen-3-8", "name": "Qwen 3.8", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=50008427", "title": "Whistle: Speech to Text in 16.9 MB", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49727511", "title": "Show HN: Swift-Qwen3.8-27B, -58.3% thinking, x1.95 speed, accuracy of xhigh", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49611128", "title": "Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49324985", "title": "Qwen 3.8 27B is excellent, but it defaults to overthinking things", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48495557", "title": "Mtplx 1.0.0", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {}, "2026-04": {}, "2026-05": {}, "2026-06": {"recommend": 1}, "2026-07": {}, "2026-08": {"recommend": 1}, "2026-09": {"recommend": 1}, "2026-10": {"neutral": 1}}, "quotes": [{"id": 800172212, "item_id": 50010763, "url": "https://news.ycombinator.com/item?id=50010763", "by": "jwr", "date": "2026-10-08", "stance": "neutral", "quote": "local Qwen 3.8 running under MTPLX", "excerpt": "People keep praising Parakeet, but I've found it to be worse than Whisper Large. Yes, it is much, much faster and smaller, but accuracy matters a lot if you are to use dictation regularly and seriously. I ended up having AI optimize Whisper Large and create a plugin for TypeWhisper, and that's what I use (feeding the results through local Qwen 3.8 running under MTPLX).", "context": "", "use_case": "running local models", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 795681441, "item_id": 49730090, "url": "https://news.ycombinator.com/item?id=49730090", "by": "bellowsgulch", "date": "2026-09-16", "stance": "warn", "quote": "I still found the general MLX MTP acceleration to be so poor", "excerpt": "I so wanted to be able to use Qwen3.8-27B locally on my fully loaded M1 Max 64GB when spider-mario recommended I use MTPLX[1], but I still found the general MLX MTP acceleration to be so poor, I'd rather just use cheap tokens from OpenCode Zen and Go. It's just too slow of a model. I know, I know there's new hardware, but my business paid like 4-5 grand for this MBP at the time, and I just don't feel the need to pay ", "context": "", "use_case": "local model acceleration", "switched_from": "", "story": "Show HN: Swift-Qwen3.8-27B, -58.3% thinking, x1.95 speed, accuracy of xhigh"}, {"id": 793811553, "item_id": 49613222, "url": "https://news.ycombinator.com/item?id=49613222", "by": "bellowsgulch", "date": "2026-09-08", "stance": "recommend", "quote": "I\u2019ll go use that now that I know.", "excerpt": "Thanks, man! I\u2019ll go use that now that I know. llama-server the last time I used it for inference with this model wasn\u2019t able to produce work fast enough to reach those numbers.", "context": "", "use_case": "faster local inference", "switched_from": "llama-server", "story": "Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses"}, {"id": 789224400, "item_id": 49326525, "url": "https://news.ycombinator.com/item?id=49326525", "by": "suoloordi", "date": "2026-08-16", "stance": "recommend", "quote": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "excerpt": "Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.", "context": "36GB MacBook Pro user", "use_case": "running Qwen locally on Mac", "switched_from": "", "story": "Qwen 3.8 27B is excellent, but it defaults to overthinking things"}, {"id": 775928928, "item_id": 48495558, "url": "https://news.ycombinator.com/item?id=48495558", "by": "rrevi", "date": "2026-06-11", "stance": "recommend", "quote": "I have been using MTPLX (replaced oMLX for me)", "excerpt": "I have been using MTPLX (replaced oMLX for me) on my M5 Pro with 64GB with Qwen3.6 27B Optimized Quality (getting about 20-25 generated tps) and along with OpenCode have been enjoying the experience very much :) HUGE shoutout to Youssof! Thank you.", "context": "M5 Pro 64GB user", "use_case": "local LLM inference", "switched_from": "oMLX", "story": "Mtplx 1.0.0"}]}