News for hackers. Stuff that matters.
Re:Microsoft AI: Seven new world-class models at scalable pr... by johnshahbazian on Wednesday November 18, 2026 @12:15PM (#50549009)
I didn't watch the keynote, but I've been testing MAI Code 2 and its been quite good for my usual work. They put it on part with Opus or Sol, and I wouldn't argue too much with that. Kudos to them for starting from scratch and getting it right.
Re:Microsoft AI: Seven new world-class models at scalable pr... by teiferer on Wednesday November 18, 2026 @12:46PM (#50549501)
I thought they were in bed with OpenAI. Did that change? I'm so not up-to-date...
Re:Microsoft AI: Seven new world-class models at scalable pr... by jhickok on Wednesday November 18, 2026 @12:55PM (#50549641)
No, for a few years now any semblance of exclusivity has been ended. They have used other model providers for a while now, including using Claude Cowork for Copilot Cowork.
Re:Microsoft AI: Seven new world-class models at scalable pr... by nicce on Wednesday November 18, 2026 @01:33PM (#50550173)
That is not completely true. They have still access to IP until 2032.

https://blogs.microsoft.com/blog/2026/04/27/the-next-phase-o...

Re:Microsoft AI: Seven new world-class models at scalable pr... by machiaweliczny on Wednesday November 18, 2026 @01:14PM (#50549896)
They have access to IP afaik
Re:Microsoft AI: Seven new world-class models at scalable pr... by mrandish on Wednesday November 18, 2026 @01:47PM (#50550377)
The OAI deal was a transitional alliance of convenience to quickly accelerate. Investing / loaning ~$10B into Open AI (much of it in compute credits) was MSFT playing catch up when they realized they were behind in the LLM race. They weren't so much worried about OAI / Ant, this was more about Google who had DeepMind and Google Brain plus significant advantage in custom TPU compute at a time when MSFT still lacked an internal frontier foundation model team.

MSFT got an IP license and a frontier model to white label along with an anchor tenant as they scaled Azure's AI compute. OAI got investment at a high valuation (good optics) and scale compute.

Re:Microsoft AI: Seven new world-class models at scalable pr... by wg0 on Wednesday November 18, 2026 @01:25PM (#50550048)
Is it lower in price then GLM 5.3 Flash and DeepSeek v4.1 Flash?
Re:Microsoft AI: Seven new world-class models at scalable pr... by jauntywundrkind on Wednesday November 18, 2026 @01:41PM (#50550300)
I wonder how much it helps that they've built their own seemingly very good AI chips, Maia 20p now Maia 300. Their blogging on these are fairly good from a spec sense (but also pretty opaque about the architectures). I don't know whether thesehttps://azure.microsoft.com/en-us/blog/deep-dive-into-azure-...

Could be mistaken but I don't think Microsoft gives direct access to customers, unlike Google tpu or aws Trainium.

Re:Microsoft AI: Seven new world-class models at scalable pr... by petercooper on Wednesday November 18, 2026 @02:00PM (#50550541)
MAI Cyber 2 Flash is listed as $0.10 input / $0.01 cached input / $0.50 output per 1M tokens which places it more on par with Luna/Haiku than DS 4.1 Flash. Didn't see the price for Code though.
Re:Microsoft AI: Seven new world-class models at scalable pr... by hiddenvulkcan on Wednesday November 18, 2026 @02:21PM (#50550780)
That is very cheap, but DeepSeek from 3rd parties goes for like $0.036 in. Source: https://openrouter.ai/deepseek/deepseek-v4.1-flash-1017
Re:Microsoft AI: Seven new world-class models at scalable pr... by qeternity on Wednesday November 18, 2026 @03:29PM (#50551701)
Lmao no it isn't. That is the cheapest price and from a single random provider.

Sure, go ahead and use them. Maybe they're serving you a decent quant. Maybe they're even serving you the real model.

The absolute cheapest provider I'd ever touch is DeepInfra (they have plenty of problems) and they are $0.09 for input.

Re:Microsoft AI: Seven new world-class models at scalable pr... by wg0 on Wednesday November 18, 2026 @03:40PM (#50551839)
These cheaper ones are usually limited on context window or token speed or latency or heavily quantised.
Re:Microsoft AI: Seven new world-class models at scalable pr... by hiddenvulkcan on Wednesday November 18, 2026 @05:25PM (#50553162)
Yeah that is right, that was the lowest not average! Even at that $0.10 is very reasonable! Tbh the quant wouldn't be an issue if openrouter didn't default you to spread across different providers, makes it hard to get consistent results for evals.
Re:Microsoft AI: Seven new world-class models at scalable pr... by ckocagil on Wednesday November 18, 2026 @07:15PM (#50554241)
Click on the provider, scroll down and you'll see effective prices. Both DeepInfra and DeepSeek are at $0.008/M input. DS Flash is unbelievably cache efficient.
Re:Microsoft AI: Seven new world-class models at scalable pr... by nicce on Wednesday November 18, 2026 @01:30PM (#50550115)
> Kudos to them for starting from scratch and getting it right.

Did they? I thought OpenAI had quite bad deal with them and wanted to get away from it. And this was one of the reasons. Microsoft got access to everything.

Re:Microsoft AI: Seven new world-class models at scalable pr... by erikpukinskis on Wednesday November 18, 2026 @01:34PM (#50550185)
Right, the “bad deal” where Microsoft gave $13 billion cash and $100B worth of free hosting to a startup that was pre-revenue and didn’t own its IP...
Re:Microsoft AI: Seven new world-class models at scalable pr... by nicce on Wednesday November 18, 2026 @03:44PM (#50551879)
Microsoft invested a lot, but they can use any novel OpenAI ever develops until 2032, and they have much more money to take any advantage of it.
Re:Microsoft AI: Seven new world-class models at scalable pr... by vorticalbox on Wednesday November 18, 2026 @03:49PM (#50551941)
I wonder if that contract forbids MS from using openAI models to develop their own.
Re:Microsoft AI: Seven new world-class models at scalable pr... by zhivota on Wednesday November 18, 2026 @06:41PM (#50553936)
Hmm so the obvious question is are these models just distillations?
Re:Microsoft AI: Seven new world-class models at scalable pr... by doomroot13 on Wednesday November 18, 2026 @07:43PM (#50554498)
Re:Microsoft AI: Seven new world-class models at scalable pr... by zhivota on Wednesday November 18, 2026 @09:46PM (#50555435)
Good read, thanks. Their description of their advantage, having control over the full pipeline including data, rings true. They are probably the best suited to compete with Google in that way.

Given the low bar Google has set for the productization of AI capability by a full stack productivity supplier, MS definitely has a chance to edge Google back out of some enterprise contracts if they do this well. Not that MS really has done an amazing job, GSuite probably wouldn't exist if they hadn't fumbled the ball so much.

Re:Microsoft AI: Seven new world-class models at scalable pr... by Degorath on Wednesday November 18, 2026 @01:57PM (#50550508)
Can you give an opinion on its communication style?
[dead]
Re:Microsoft AI: Seven new world-class models at scalable pr... by Degorath on Wednesday November 18, 2026 @05:40PM (#50553374)
Very cool, thanks for the examples!

If you had to compare it to sol or opus, is it more easy to back-and-forth with? Imo opus devolves into word salad quite easily.

Re:Microsoft AI: Seven new world-class models at scalable pr... by jwpapi on Wednesday November 18, 2026 @03:47PM (#50551913)
How do you test it ? Is it on your Azure Foundry?
Re:Microsoft AI: Seven new world-class models at scalable pr... by greggsy on Wednesday November 18, 2026 @04:43PM (#50552540)
For me, GitHub Copilot via VS Code
Re:Microsoft AI: Seven new world-class models at scalable pr... by johnshahbazian on Wednesday November 18, 2026 @04:44PM (#50552549)
Visual Studio's built-in Github Copilot. Though they have a desktop app of their own now.
Re:Microsoft AI: Seven new world-class models at scalable pr... by cicko on Wednesday November 18, 2026 @06:16PM (#50553728)
I thought they were focused on a VS Code extension. Thanks for the info.
Re:Microsoft AI: Seven new world-class models at scalable pr... by doomroot13 on Wednesday November 18, 2026 @07:44PM (#50554513)
They will also soon be releasing on OpenRouter, Vercel, Hermes, OpenCode, and Kilo.