{"author":"guybedo","children":[{"author":"trehcrob","children":[],"created_at":"2026-10-21T19:52:41.000Z","created_at_i":1792612361,"id":50186646,"options":[],"parent_id":50183678,"points":null,"story_id":50183678,"text":"Congrats on the release, hope the next iteration outperforms the latest Qwen model.","title":null,"type":"comment","url":null},{"author":"kamranjon","children":[],"created_at":"2026-10-21T23:01:50.000Z","created_at_i":1792623710,"id":50188873,"options":[],"parent_id":50183678,"points":null,"story_id":50183678,"text":"Pretty amazing for being trained from scratch, to have ~Qwen 3.8 27b quality with such an increase in tok&#x2F;s performance. Excited to try it out and see how snappy it is.","title":null,"type":"comment","url":null},{"author":"nvtop","children":[{"author":"dofm","children":[],"created_at":"2026-10-22T20:51:57.000Z","created_at_i":1792702317,"id":50200845,"options":[],"parent_id":50200303,"points":null,"story_id":50183678,"text":"I&#x27;m just about to test this one out (M1 Max).<p>I really like Muse Glimmer, and on low reasoning level it&#x27;s both helpful and quite quick.<p>But I won&#x27;t turn down a model that is faster.<p>Nemotron Bolt appears to be really quite good but unfortunately it is quite slow on macOS&#x2F;Metal in llama.cpp at the moment because of lack of support for a tensor type (or something\u2026 not sure I understand).","title":null,"type":"comment","url":null}],"created_at":"2026-10-22T20:08:51.000Z","created_at_i":1792699731,"id":50200303,"options":[],"parent_id":50183678,"points":null,"story_id":50183678,"text":"- Better than Gemma 4 and Muse Glimmer on SWE benchmarks<p>- 125 tokens per second on an M4 Pro MacBook Pro<p>- 512k tokens context window<p>- 40B total, 2.5B active parameters mixture-of-experts<p>Pretty cool release!","title":null,"type":"comment","url":null},{"author":"aistripes","children":[],"created_at":"2026-10-22T22:21:35.000Z","created_at_i":1792707695,"id":50201889,"options":[],"parent_id":50183678,"points":null,"story_id":50183678,"text":"Numbers seem good but can\u2019t really be impressed on real world test <a href=\"https:&#x2F;&#x2F;hypeorcode.com&#x2F;models&#x2F;flint&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;hypeorcode.com&#x2F;models&#x2F;flint&#x2F;</a>","title":null,"type":"comment","url":null},{"author":"nittanymount","children":[],"created_at":"2026-10-23T21:21:23.000Z","created_at_i":1792790483,"id":50215526,"options":[],"parent_id":50183678,"points":null,"story_id":50183678,"text":"not as good as qwen3.8 yet, even the chart says good numbers, but","title":null,"type":"comment","url":null}],"created_at":"2026-10-21T16:31:03.000Z","created_at_i":1792600263,"id":50183678,"options":[],"parent_id":null,"points":29,"story_id":50183678,"text":null,"title":"Recursive Introduces Flint Model","type":"story","url":"https://recursive.com/updates/introducing-recursive-flint"}