{"exhaustive":{"nbHits":false,"typo":false},"exhaustiveNbHits":false,"exhaustiveTypo":false,"hits":[{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"lisajaloza"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["story","author_lisajaloza","story_50024913"],"author":"lisajaloza","children":[50025130,50025415,50025687,50025708,50025729,50025768,50025796,50026346,50026865,50027142,50027180,50027424,50027810,50027896,50028142,50028380,50029354,50029699,50031428,50035989],"created_at":"2026-10-09T18:38:44Z","created_at_i":1791571124,"num_comments":66,"objectID":"50024913","points":201,"story_id":50024913,"title":"Microsoft-Decision-1, our model for fast decision-making","updated_at":"2026-11-21T21:18:57Z","url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gmays"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft Decision-1</em>: 9B Qwen Model Launch Explained"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://shattered.io/<em>microsoft-decision-1</em>-qwen3-5-9b-model-2026/"}},"_tags":["story","author_gmays","story_50238620"],"author":"gmays","children":[50239324,50239355,50239429,50239483,50241140,50241734],"created_at":"2026-10-26T03:12:05Z","created_at_i":1792984325,"num_comments":32,"objectID":"50238620","points":72,"story_id":50238620,"title":"Microsoft Decision-1: 9B Qwen Model Launch Explained","updated_at":"2026-11-04T11:22:37Z","url":"https://shattered.io/microsoft-decision-1-qwen3-5-9b-model-2026/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"lisajaloza"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>.1"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://commandline.microsoft.com/microsoft-decision-model-ga/"}},"_tags":["story","author_lisajaloza","story_50535763"],"author":"lisajaloza","children":[50537004],"created_at":"2026-11-17T18:17:28Z","created_at_i":1794939448,"num_comments":1,"objectID":"50535763","points":8,"story_id":50535763,"title":"Microsoft-Decision-1.1","updated_at":"2026-11-18T00:11:48Z","url":"https://commandline.microsoft.com/microsoft-decision-model-ga/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mrkn1"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Microsoft Launched <em>Microsoft-Decision-1</em> decision model"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://twitter.com/satyanadella/status/2108627923888754862"}},"_tags":["story","author_mrkn1","story_50025521"],"author":"mrkn1","children":[50025851],"created_at":"2026-10-09T19:28:12Z","created_at_i":1791574092,"num_comments":2,"objectID":"50025521","points":2,"story_id":50025521,"title":"Microsoft Launched Microsoft-Decision-1 decision model","updated_at":"2026-10-09T20:06:59Z","url":"https://twitter.com/satyanadella/status/2108627923888754862"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gmays"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>.1 is now GA for fast, low-cost decision making"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://commandline.microsoft.com/microsoft-decision-model-ga/"}},"_tags":["story","author_gmays","story_50555832"],"author":"gmays","created_at":"2026-11-19T03:46:29Z","created_at_i":1795059989,"num_comments":0,"objectID":"50555832","points":2,"story_id":50555832,"title":"Microsoft-Decision-1.1 is now GA for fast, low-cost decision making","updated_at":"2026-11-19T05:40:53Z","url":"https://commandline.microsoft.com/microsoft-decision-model-ga/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ycosynot"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Using <em>Microsoft-Decision-1</em> (out yesterday) to fill 3D indoor levels with assets"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/yeme-oss/LevelDez/"}},"_tags":["story","author_ycosynot","story_50045309"],"author":"ycosynot","children":[50045310],"created_at":"2026-10-11T17:01:12Z","created_at_i":1791738072,"num_comments":0,"objectID":"50045309","points":1,"story_id":50045309,"title":"Using Microsoft-Decision-1 (out yesterday) to fill 3D indoor levels with assets","updated_at":"2026-10-11T17:02:19Z","url":"https://github.com/yeme-oss/LevelDez/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jmoore15"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Hi,<p>John here from the team that built <em>Microsoft Decision 1</em> and 1.1.<p><em>Microsoft Decision 1</em>.1 is on the Artificial Analysis Decision Index v2.1 leaderboard, where score 68, above Gemini Decision 1, GPT-6 Luna Decisions, and Jev 1.14. [1]<p>We also compare ourselves to other decision models in our blog post on Microsoft's Command Line blog from November 17 [2]. We name all of the public benchmarks we evaluated against in the post, and provide a breakdown by category.<p>[1] <a href=\"https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index\" rel=\"nofollow\">https://artificialanalysis.ai/evaluations/artificial-analysi...</a>\n[2] <a href=\"https://commandline.microsoft.com/microsoft-decision-model-ga/\" rel=\"nofollow\">https://commandline.microsoft.com/microsoft-decision-model-g...</a>"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Seven new models at scalable price-performance"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://microsoft.ai/news/microsoft-ignite-2026/"}},"_tags":["comment","author_jmoore15","story_50548021"],"author":"jmoore15","children":[50551806],"comment_text":"Hi,<p>John here from the team that built Microsoft Decision 1 and 1.1.<p>Microsoft Decision 1.1 is on the Artificial Analysis Decision Index v2.1 leaderboard, where score 68, above Gemini Decision 1, GPT-6 Luna Decisions, and Jev 1.14. [1]<p>We also compare ourselves to other decision models in our blog post on Microsoft&#x27;s Command Line blog from November 17 [2]. We name all of the public benchmarks we evaluated against in the post, and provide a breakdown by category.<p>[1] <a href=\"https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;evaluations&#x2F;artificial-analysis-decision-index\" rel=\"nofollow\">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;evaluations&#x2F;artificial-analysi...</a>\n[2] <a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-model-ga&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-model-g...</a>","created_at":"2026-11-18T19:51:23Z","created_at_i":1795031483,"objectID":"50551205","parent_id":50550492,"story_id":50548021,"story_title":"Seven new models at scalable price-performance","story_url":"https://microsoft.ai/news/microsoft-ignite-2026/","updated_at":"2026-11-19T06:02:38Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jmoore15"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Hi,<p>John here from the team that built <em>Microsoft Decision 1</em> and 1.1.<p>We compare ourselves to other decision models in our blog post on Microsoft's Command Line blog from November 17 [1]. We name all of the public benchmarks we evaluated against in the post, and provide a breakdown by category.<p>We also are on the Artificial Analysis leaderboard, which is an independent third party assessment. On AA, we score 68, above Gemini Decision 1, GPT-6 Luna Decisions, and Jev 1.14. [2]<p>Our previous MD1 model from October is also benchmarked on JevBench, another independent benchmark for decision models. [3] On JevBench, we are 9th place for API models, ahead of Gemini Decision 1 and Jev 1.14. We look forward to seeing how the new MD1.1 scores on JevBench.<p>With regards to cost: Our models are the same price as Jev on a per-token basis. However, on both JevBench and Artificial Analysis benchmarks, our models use fewer tokens per task than Jev. For example, our cost score for MD1.1 on Artificial Analysis at $140 per 1M tasks compared to Jev's $156 and Gemini's $345. There are similar stats for latency, where our new MD1.1 model is 2x faster than our previous one.<p>[1] <a href=\"https://commandline.microsoft.com/microsoft-decision-model-ga/\" rel=\"nofollow\">https://commandline.microsoft.com/microsoft-decision-model-g...</a>\n[2] <a href=\"https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index\" rel=\"nofollow\">https://artificialanalysis.ai/evaluations/artificial-analysi...</a>\n[3] <a href=\"https://benchmarkheaven.com/jev-models/api\" rel=\"nofollow\">https://benchmarkheaven.com/jev-models/api</a>"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Seven new models at scalable price-performance"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://microsoft.ai/news/microsoft-ignite-2026/"}},"_tags":["comment","author_jmoore15","story_50548021"],"author":"jmoore15","comment_text":"Hi,<p>John here from the team that built Microsoft Decision 1 and 1.1.<p>We compare ourselves to other decision models in our blog post on Microsoft&#x27;s Command Line blog from November 17 [1]. We name all of the public benchmarks we evaluated against in the post, and provide a breakdown by category.<p>We also are on the Artificial Analysis leaderboard, which is an independent third party assessment. On AA, we score 68, above Gemini Decision 1, GPT-6 Luna Decisions, and Jev 1.14. [2]<p>Our previous MD1 model from October is also benchmarked on JevBench, another independent benchmark for decision models. [3] On JevBench, we are 9th place for API models, ahead of Gemini Decision 1 and Jev 1.14. We look forward to seeing how the new MD1.1 scores on JevBench.<p>With regards to cost: Our models are the same price as Jev on a per-token basis. However, on both JevBench and Artificial Analysis benchmarks, our models use fewer tokens per task than Jev. For example, our cost score for MD1.1 on Artificial Analysis at $140 per 1M tasks compared to Jev&#x27;s $156 and Gemini&#x27;s $345. There are similar stats for latency, where our new MD1.1 model is 2x faster than our previous one.<p>[1] <a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-model-ga&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-model-g...</a>\n[2] <a href=\"https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;evaluations&#x2F;artificial-analysis-decision-index\" rel=\"nofollow\">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;evaluations&#x2F;artificial-analysi...</a>\n[3] <a href=\"https:&#x2F;&#x2F;benchmarkheaven.com&#x2F;jev-models&#x2F;api\" rel=\"nofollow\">https:&#x2F;&#x2F;benchmarkheaven.com&#x2F;jev-models&#x2F;api</a>","created_at":"2026-11-18T19:32:02Z","created_at_i":1795030322,"objectID":"50550940","parent_id":50548399,"story_id":50548021,"story_title":"Seven new models at scalable price-performance","story_url":"https://microsoft.ai/news/microsoft-ignite-2026/","updated_at":"2026-11-19T02:51:07Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"usmank06"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"When Jev came out, it was a big thing, but many other similar decision models came out pretty quickly. I'm currently using Jev for a side project, and I wanted to know if the actual decision training does anything, or if it's pretty much the same as an LLM, just answering in one token.<p>Two of them (Elephant 1 and <em>Microsoft-Decision-1</em>) are built on Qwen3.5-9B, so I ran them against plain Qwen3.5-9B answering in one token, and the same Qwen thinking first. On the reasoning questions: one token got 0.53, the decision models 0.69 and 0.63, thinking got 0.99. So the training helps a bit, but it doesn't bring the reasoning back, it just speeds things up and reduces cost.<p>Everything's in the link and the repo, if y'all have any questions I'm happy to answer them :)"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"[dead]"}},"_tags":["comment","author_usmank06","story_50493665"],"author":"usmank06","comment_text":"When Jev came out, it was a big thing, but many other similar decision models came out pretty quickly. I&#x27;m currently using Jev for a side project, and I wanted to know if the actual decision training does anything, or if it&#x27;s pretty much the same as an LLM, just answering in one token.<p>Two of them (Elephant 1 and Microsoft-Decision-1) are built on Qwen3.5-9B, so I ran them against plain Qwen3.5-9B answering in one token, and the same Qwen thinking first. On the reasoning questions: one token got 0.53, the decision models 0.69 and 0.63, thinking got 0.99. So the training helps a bit, but it doesn&#x27;t bring the reasoning back, it just speeds things up and reduces cost.<p>Everything&#x27;s in the link and the repo, if y&#x27;all have any questions I&#x27;m happy to answer them :)","created_at":"2026-11-14T03:49:50Z","created_at_i":1794628190,"objectID":"50493666","parent_id":50493665,"story_id":50493665,"story_title":"[dead]","updated_at":"2026-11-14T03:52:07Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"CharlieDigital"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Many F500 are already using Chinese models or models derived from them through American providers.  Even Microsoft's Decision-1 Jev-like:<p><a href=\"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/\" rel=\"nofollow\">https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model...</a><p><pre><code>    &gt; To build <em>Microsoft-Decision-1</em>, we post trained Qwen3.5-9B for fast, single-pass decision scoring and will soon rebase it on other models, including Microsoft AI (MAI) and OpenAI.\n</code></pre>\nCursor's models most famously are built on top of Kimi:<p><a href=\"https://techcrunch.com/2026/03/22/cursor-admits-its-new-coding-model-was-built-on-top-of-moonshot-ais-kimi/\" rel=\"nofollow\">https://techcrunch.com/2026/03/22/cursor-admits-its-new-codi...</a><p><pre><code>    &gt; Cursor\u2019s vice president of developer education Lee Robinson soon acknowledged, \u201cYep, Composer 2 started from an open-source base!\u201d But he said, \u201cOnly ~1/4 of the compute spent on the final model came from the base, the rest is from our training.\u201d As a result, he said Composer 2\u2019s performance on various benchmarks is \u201cvery different\u201d from Kimi\u2019s.\n</code></pre>\nSo yeah, Chinese open weight models are already in F500."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Anthropic IPO buyers see valuation as low as $1.1T"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.bloomberg.com/news/articles/2026-11-06/anthropic-ipo-valuation-expectations-pressured-after-ai-rout"}},"_tags":["comment","author_CharlieDigital","story_50394906"],"author":"CharlieDigital","comment_text":"Many F500 are already using Chinese models or models derived from them through American providers.  Even Microsoft&#x27;s Decision-1 Jev-like:<p><a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model-foundry&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model...</a><p><pre><code>    &gt; To build Microsoft-Decision-1, we post trained Qwen3.5-9B for fast, single-pass decision scoring and will soon rebase it on other models, including Microsoft AI (MAI) and OpenAI.\n</code></pre>\nCursor&#x27;s models most famously are built on top of Kimi:<p><a href=\"https:&#x2F;&#x2F;techcrunch.com&#x2F;2026&#x2F;03&#x2F;22&#x2F;cursor-admits-its-new-coding-model-was-built-on-top-of-moonshot-ais-kimi&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;techcrunch.com&#x2F;2026&#x2F;03&#x2F;22&#x2F;cursor-admits-its-new-codi...</a><p><pre><code>    &gt; Cursor\u2019s vice president of developer education Lee Robinson soon acknowledged, \u201cYep, Composer 2 started from an open-source base!\u201d But he said, \u201cOnly ~1&#x2F;4 of the compute spent on the final model came from the base, the rest is from our training.\u201d As a result, he said Composer 2\u2019s performance on various benchmarks is \u201cvery different\u201d from Kimi\u2019s.\n</code></pre>\nSo yeah, Chinese open weight models are already in F500.","created_at":"2026-11-06T19:11:00Z","created_at_i":1793992260,"objectID":"50396893","parent_id":50396447,"story_id":50394906,"story_title":"Anthropic IPO buyers see valuation as low as $1.1T","story_url":"https://www.bloomberg.com/news/articles/2026-11-06/anthropic-ipo-valuation-expectations-pressured-after-ai-rout","updated_at":"2026-11-07T15:33:00Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ricardobeat"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"&gt; a model that refuses to write you a paragraph<p>AI slop reporting. Actual news: <a href=\"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/\" rel=\"nofollow\">https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model...</a>"},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft Decision-1</em>: 9B Qwen Model Launch Explained"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://shattered.io/<em>microsoft-decision-1</em>-qwen3-5-9b-model-2026/"}},"_tags":["comment","author_ricardobeat","story_50238620"],"author":"ricardobeat","comment_text":"&gt; a model that refuses to write you a paragraph<p>AI slop reporting. Actual news: <a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model-foundry&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model...</a>","created_at":"2026-10-26T10:45:11Z","created_at_i":1793011511,"objectID":"50241734","parent_id":50238620,"story_id":50238620,"story_title":"Microsoft Decision-1: 9B Qwen Model Launch Explained","story_url":"https://shattered.io/microsoft-decision-1-qwen3-5-9b-model-2026/","updated_at":"2026-10-26T11:10:19Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"joelsstafford"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Let's test it in a real-world trading scenario. Jev 1.13 had quite good results, only beaten by <em>Microsoft-Decision-1</em>: <a href=\"https://pilabs.space/trading-arena\" rel=\"nofollow\">https://pilabs.space/trading-arena</a>"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"We\u2019re releasing Jev 1.14.0 along with experimental features for dreamers"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://typesafe.ai/blog/jev-1-14"}},"_tags":["comment","author_joelsstafford","story_50095353"],"author":"joelsstafford","children":[50101054],"comment_text":"Let&#x27;s test it in a real-world trading scenario. Jev 1.13 had quite good results, only beaten by Microsoft-Decision-1: <a href=\"https:&#x2F;&#x2F;pilabs.space&#x2F;trading-arena\" rel=\"nofollow\">https:&#x2F;&#x2F;pilabs.space&#x2F;trading-arena</a>","created_at":"2026-10-15T08:42:06Z","created_at_i":1792053726,"objectID":"50100943","parent_id":50095353,"story_id":50095353,"story_title":"We\u2019re releasing Jev 1.14.0 along with experimental features for dreamers","story_url":"https://typesafe.ai/blog/jev-1-14","updated_at":"2026-10-15T10:51:31Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"felix089"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Good timing with our pac-man benchmark for decision models, now also including <em>Microsoft Decision-1</em>: <a href=\"https://news.ycombinator.com/item?id=50007993\">https://news.ycombinator.com/item?id=50007993</a>"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Typesafe AI raises $870M at $7.5B"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://typesafe.ai/blog/series-ai"}},"_tags":["comment","author_felix089","story_50023450"],"author":"felix089","comment_text":"Good timing with our pac-man benchmark for decision models, now also including Microsoft Decision-1: <a href=\"https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=50007993\">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=50007993</a>","created_at":"2026-10-10T13:44:33Z","created_at_i":1791639873,"objectID":"50032919","parent_id":50023450,"story_id":50023450,"story_title":"Typesafe AI raises $870M at $7.5B","story_url":"https://typesafe.ai/blog/series-ai","updated_at":"2026-10-10T13:50:01Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jitbit"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"They do mention it in the article<p>\u201cTo build <em>Microsoft-Decision-1</em>, we post trained Qwen3.5-9B\u201d"},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["comment","author_jitbit","story_50024913"],"author":"jitbit","comment_text":"They do mention it in the article<p>\u201cTo build Microsoft-Decision-1, we post trained Qwen3.5-9B\u201d","created_at":"2026-10-10T12:57:12Z","created_at_i":1791637032,"objectID":"50032479","parent_id":50028401,"story_id":50024913,"story_title":"Microsoft-Decision-1, our model for fast decision-making","story_url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/","updated_at":"2026-10-10T18:29:30Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hulitu"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"&gt; <em>Microsoft-Decision-1</em>, our model for fast decision-making<p>I'm still waiting for the Microsoft Vacuum Cleaner. /s"},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["comment","author_hulitu","story_50024913"],"author":"hulitu","comment_text":"&gt; Microsoft-Decision-1, our model for fast decision-making<p>I&#x27;m still waiting for the Microsoft Vacuum Cleaner. &#x2F;s","created_at":"2026-10-10T10:21:55Z","created_at_i":1791627715,"objectID":"50031428","parent_id":50024913,"story_id":50024913,"story_title":"Microsoft-Decision-1, our model for fast decision-making","story_url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/","updated_at":"2026-10-10T10:25:15Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"pseudotensor"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"Update: Microsoft just announced <em>Microsoft-Decision-1</em> (a post-trained Qwen3.5-9B) and benchmarked it against H2O-Lightning-4B, among others: <a href=\"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/\" rel=\"nofollow\">https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model...</a><p>One note on their speed comparison: the JevBench board shows self-hosted models with an adjusted latency of &quot;2x + 0.15 s (assumption, not measured)&quot;. Our measured p50 is 29 ms; the adjusted figure is 0.21 s. They report 85 ms for theirs."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Typesafe AI raises $870M at $7.5B"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://typesafe.ai/blog/series-ai"}},"_tags":["comment","author_pseudotensor","story_50023450"],"author":"pseudotensor","comment_text":"Update: Microsoft just announced Microsoft-Decision-1 (a post-trained Qwen3.5-9B) and benchmarked it against H2O-Lightning-4B, among others: <a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model-foundry&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model...</a><p>One note on their speed comparison: the JevBench board shows self-hosted models with an adjusted latency of &quot;2x + 0.15 s (assumption, not measured)&quot;. Our measured p50 is 29 ms; the adjusted figure is 0.21 s. They report 85 ms for theirs.","created_at":"2026-10-09T21:45:12Z","created_at_i":1791582312,"objectID":"50026966","parent_id":50026118,"story_id":50023450,"story_title":"Typesafe AI raises $870M at $7.5B","story_url":"https://typesafe.ai/blog/series-ai","updated_at":"2026-10-10T10:22:44Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jongalloway2"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<a href=\"https://ai.azure.com/catalog/models/Microsoft-Decision-1\" rel=\"nofollow\">https://ai.azure.com/catalog/models/<em>Microsoft-Decision-1</em></a>"},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["comment","author_jongalloway2","story_50024913"],"author":"jongalloway2","children":[50026273],"comment_text":"<a href=\"https:&#x2F;&#x2F;ai.azure.com&#x2F;catalog&#x2F;models&#x2F;Microsoft-Decision-1\" rel=\"nofollow\">https:&#x2F;&#x2F;ai.azure.com&#x2F;catalog&#x2F;models&#x2F;Microsoft-Decision-1</a>","created_at":"2026-10-09T20:17:22Z","created_at_i":1791577042,"objectID":"50026062","parent_id":50025768,"story_id":50024913,"story_title":"Microsoft-Decision-1, our model for fast decision-making","story_url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/","updated_at":"2026-10-10T07:36:44Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"simonw"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"&gt; To build <em>Microsoft-Decision-1</em>, we post trained Qwen3.5-9B for fast, single-pass decision scoring and will soon rebase it on other models, including Microsoft AI (MAI) and OpenAI.<p>I guess the fear of Chinese models is finally subsiding."},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["comment","author_simonw","story_50024913"],"author":"simonw","children":[50025988,50100581],"comment_text":"&gt; To build Microsoft-Decision-1, we post trained Qwen3.5-9B for fast, single-pass decision scoring and will soon rebase it on other models, including Microsoft AI (MAI) and OpenAI.<p>I guess the fear of Chinese models is finally subsiding.","created_at":"2026-10-09T19:44:22Z","created_at_i":1791575062,"objectID":"50025708","parent_id":50024913,"story_id":50024913,"story_title":"Microsoft-Decision-1, our model for fast decision-making","story_url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/","updated_at":"2026-10-15T07:56:15Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ampersandwhich"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"I hope they can finally make my &quot;Copilot+ PC&quot; infer things locally that are actually useful. Phi Silica for Advanced Paste was a good start, if a bit late. If they got their act together, <em>Microsoft-Decision-1</em> could have some local potential. Their track record leaves me with some reservations."},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"<em>Microsoft-Decision-1</em>, our model for fast decision-making"},"story_url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model-foundry/"}},"_tags":["comment","author_ampersandwhich","story_50024913"],"author":"ampersandwhich","children":[50025727],"comment_text":"I hope they can finally make my &quot;Copilot+ PC&quot; infer things locally that are actually useful. Phi Silica for Advanced Paste was a good start, if a bit late. If they got their act together, Microsoft-Decision-1 could have some local potential. Their track record leaves me with some reservations.","created_at":"2026-10-09T19:29:18Z","created_at_i":1791574158,"objectID":"50025534","parent_id":50025130,"story_id":50024913,"story_title":"Microsoft-Decision-1, our model for fast decision-making","story_url":"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/","updated_at":"2026-10-10T02:02:43Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"armcat"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["microsoft","decision","1"],"value":"It's weird because two days after Jev was released there were a dozen decision models, a week later there are several dozen, mostly open source, OpenAI's own Decisions API [1] beats it, and you can easily finetune your own [2]. But as others have pointed out, this doesn't matter.<p>EDIT: As I wrote this Microsoft just released their own Decision-1 model [3].<p>[1] <a href=\"https://developers.openai.com/api/docs/guides/decisions\" rel=\"nofollow\">https://developers.openai.com/api/docs/guides/decisions</a><p>[2] <a href=\"https://unsloth.ai/docs/basics/train-your-own-decision-model-with-unsloth\">https://unsloth.ai/docs/basics/train-your-own-decision-model...</a><p>[3] <a href=\"https://commandline.microsoft.com/microsoft-decision-1-model-foundry/\" rel=\"nofollow\">https://commandline.microsoft.com/<em>microsoft-decision-1</em>-model...</a>"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Typesafe AI raises $870M at $7.5B"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://typesafe.ai/blog/series-ai"}},"_tags":["comment","author_armcat","story_50023450"],"author":"armcat","children":[50025032,50025079,50025104,50025155,50025216,50025244,50025397,50025522,50025533,50025599,50025618,50025656,50025811,50026008,50026128,50026329,50026376,50026956,50027019,50027489,50027570,50027807,50028525,50029104,50029194,50030325,50030503,50031096],"comment_text":"It&#x27;s weird because two days after Jev was released there were a dozen decision models, a week later there are several dozen, mostly open source, OpenAI&#x27;s own Decisions API [1] beats it, and you can easily finetune your own [2]. But as others have pointed out, this doesn&#x27;t matter.<p>EDIT: As I wrote this Microsoft just released their own Decision-1 model [3].<p>[1] <a href=\"https:&#x2F;&#x2F;developers.openai.com&#x2F;api&#x2F;docs&#x2F;guides&#x2F;decisions\" rel=\"nofollow\">https:&#x2F;&#x2F;developers.openai.com&#x2F;api&#x2F;docs&#x2F;guides&#x2F;decisions</a><p>[2] <a href=\"https:&#x2F;&#x2F;unsloth.ai&#x2F;docs&#x2F;basics&#x2F;train-your-own-decision-model-with-unsloth\">https:&#x2F;&#x2F;unsloth.ai&#x2F;docs&#x2F;basics&#x2F;train-your-own-decision-model...</a><p>[3] <a href=\"https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model-foundry&#x2F;\" rel=\"nofollow\">https:&#x2F;&#x2F;commandline.microsoft.com&#x2F;microsoft-decision-1-model...</a>","created_at":"2026-10-09T18:41:42Z","created_at_i":1791571302,"objectID":"50024954","parent_id":50023450,"story_id":50023450,"story_title":"Typesafe AI raises $870M at $7.5B","story_url":"https://typesafe.ai/blog/series-ai","updated_at":"2026-11-09T17:32:22Z"}],"hitsPerPage":20,"nbHits":20,"nbPages":1,"page":0,"params":"query=Microsoft-Decision-1&advancedSyntax=true&analyticsTags=backend","processingTimeMS":57,"processingTimingsMS":{"_request":{"queue":21,"roundTrip":17},"afterFetch":{"merge":{"mergeLoop":{"prepareNextHit":2,"total":2},"total":2},"total":2},"fetch":{"query":7,"scanning":46,"total":54},"total":57},"query":"Microsoft-Decision-1","serverTimeMS":79}
