https://t.co/PLL0hNPbdv

\n
@teortaxesTex:\nGood guy DeepSeek gives us accelerated models\nThe most interesting one here is Gemma4-12B, I presume vision included. Might be the best local model in its weight class now, by some margin\nQwen 3.5 not included because DS[park] doesn't do linear attention I guess https://t.co/lKtcGBQH3w
\n

$\"\"$

","freeContentEncoded","textContent","published","strategy","unknown","locked","deletedAt","isSyncedFromOwnBlog","original","duplicatesCount","duplicateVersion","originalArticleId","errorsCount","appearance","twitter:repost","color","backgroundColor","shortSummary","summary","bullets","persons","places","words","sentiment","subject","meta","eventId","c8b53b39d28376923b7944862508c5ca3715814cd961071d33bae865b8615594","nostrRootEvent","event",{"_19":214,"_218":219,"_220":110,"_68":221,"_232":26,"_184":233,"_234":235},"sig","ca6b4ef0d4f87f9849cb6560eae007e163d26fc6d57ee96af4aa9b8a77dc27cc3ca4d81b2ec2fc3225c559941df899e28e615a72ad3a07930ce7405d8d60d5cb","kind",[222,223,227,230],[211,172],[224,225,40,226],"a","30023:983dd797fe72ad62fc8e3ece7eed1f126550972885c8f835d9af5528dad655a3:019f12d8-0362-0202-716c-0cef24509712","mention",[228,229],"client","Layer3.press",[231,26,40],"p","pubkey","https://t.co/PLL0hNPbdv\n\n\n\nnostr:naddr1qvzqqqr4gupzpxpa67tluu4dvt7gu0kw0mk37yn92ztj3pwglq6ant649rddv4drqqjrqvfevccnyepc95crxd3j95cryvpj95mnzdnr95cxxetxxg6r2vpexucnykf8yuj ","created_at",1782681008,"authorNPub","markdown","https://t.co/PLL0hNPbdv\n\n> @teortaxesTex:\n> Good guy DeepSeek gives us accelerated models\n> The most interesting one here is Gemma4-12B, I presume vision included. Might be the best local model in its weight class now, by some margin\n> Qwen 3.5 not included because DS[park] doesn't do linear attention I guess https://t.co/lKtcGBQH3w\n\n\n \n![](https://pbs.twimg.com/media/HL7Ig98WMAEyMK8.png)","freeMarkdown","kind1EventId","kind30023EventId","d9c877d1d52b8310546c4f93fb7854be7555fac274718ca52a871b0379bc8015","kind30023",{"_19":242,"_218":245,"_220":246,"_68":247,"_232":26,"_184":259,"_234":235},"8689327a00d6fe2dbca41a00f1037ce32c673b2986c6001d7d8cf0c711995d667336b9275bf5267487d0d9b7e3c58eff1c94eafe31fa7a2e728e46ceaa1b620d",30023,[248,251,252,255,257,258],[249,250],"r","https://bender.layer3.press/articles/019f12d8-0362-0202-716c-0cef24509712",[31,172],[253,254],"published_at","1782681008",[256,166],"d",[228,229],[231,26,40],"https://t.co/PLL0hNPbdv\n\n> @teortaxesTex:\n> Good guy DeepSeek gives us accelerated models\n> The most interesting one here is Gemma4-12B, I presume vision included. Might be the best local model in its weight class now, by some margin\n> Qwen 3.5 not included because DS[park] doesn't do linear attention I guess https://t.co/lKtcGBQH3w\n\n\n \n![](https://pbs.twimg.com/media/HL7Ig98WMAEyMK8.png)\n\nhttps://x.com/ClementDelangue/status/2071340393447272457","publishedToNostr","htmlContentKind30023","pubDate","2026-06-28T21:10:08.000Z","modifiedAt","2026-06-29T10:06:15.391Z","storyId","likes","zaps","reposts","comments","tweets",[],"relatedArticles",[],"actionData","errors"]