Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

MAI-Code-1-Flash

microsoft.ai

Post reply on HN
Print view View original

171–180 of 297 posts

Re: MAI-Code-1-Flash

#171

Post by lostmsu »

minraws wrote: post #62
giancarlostoro wrote: post #57

Earlier quoted context omitted.

The take away is that this model is a smaller model that competes with Haiku, I would hope they come out with a "Sonnet" competing model, then Opus. I have been wondering why Microsoft is kind of "sleeping" on offering models they themselves have made on Copilot, maybe it was part of their deal with OpenAI? Not sure.

They did release, MAI-Thinking-1 to compete with Sonnet. Totally not sure why that isn't at the top here.

Compete? It is behind Kimi K2.6, which is in turn away behind Sonnet.

Re: MAI-Code-1-Flash

#172

Post by eli »

stefan_ wrote: post #116
alkonaut wrote: post #112

Earlier quoted context omitted.

Won’t (presumably) all the market actors converge on similar pricing? If OpenAI stopped operating on subsidies and charge the true costs and their most token hungry customers are the ones that switch to Anthropic and others, then their pricing model switch will also be around the corner. Unless of course we’re thinking Copilot will be more expensive than others longer term. But is that a reasonable assumption?

Anthropic & co charge API users much more, not least to demolish the middlemen low-effort plays like Cursor and Copilot. To not own the model is not viable in 2026.

I think it’s more correct to say they charge subscription users much much less. I assume less even than the cost of providing the inference, if you actually are using it.

Re: MAI-Code-1-Flash

#173

Post by not_a_bot_4sho »

singhkays wrote: post #130
Handy-Man wrote: post #122

Earlier quoted context omitted.

That's neither Microsoft nor Anthropic design. It's from their acquisition of Inflection AI. Even Copilot mobile app design is basically what was Inflection's design

I've always wondered where Consumer CoPilot's design language was from. If you watch the Build keynote with Satya, you'll notice that the design of the slides changed to Serif typography and warmer colors when Mustafa/Microsoft AI segment came on which was completely different from the rest of the keynote. Now it makes sense!

Tangentially, I've seen a couple people use "CoPilot" here instead of "Copilot" (the actual product name).

Where does the Pascal case inspired variant come from? Is it a reference to something? Is it like "M$" was used back in the days?

Re: MAI-Code-1-Flash

#174

Post by GaryBluto »

papascrubs wrote: post #168
GaryBluto wrote: post #101

Earlier quoted context omitted.

Almost exactly the same story here. I've also had little to no refusals from DeepSeek, with it's Chinese values meaning substantially less friction when it comes to things like reverse engineering, finding copyrighted files, working with dubiously-sourced source code, et cetera. I don't think I'd go back to Copilot even if they dropped prices by 90%.

Are you purchasing directly from DeepSeek? Any concerns as far as privacy or data protection?

Using OpenRouter, going to migrate to DeepSeek's official API soon. I'm not using it for anything commercial or for private data so I have no privacy qualms.

Re: MAI-Code-1-Flash

#175

Post by eli »

bel8 wrote: post #82

It's a start and I welcome competition but I don't think I ever used small cloud models like Haiku 4.5. They are cute but for serious coding they tend to waste your expensive time. And this certainly wont bring me back to GitHub Copilot which I cancelled yesterday. GitHub Copilot had competitive pricing until yesterday when they changed from per-request to one of the most expensive per-token quotas. Seriously, take a…

Makes sense as part of a larger coding workflow, especially if it’s fast. Using a trillion parameter model to figure out how to call a targeted edit tool or generate a commit message is a waste. Also narrow tasks like “make the background darker” or “rename this function and update callers”

Re: MAI-Code-1-Flash

#176

Post by bbstats »

bel8 wrote: post #82

It's a start and I welcome competition but I don't think I ever used small cloud models like Haiku 4.5. They are cute but for serious coding they tend to waste your expensive time. And this certainly wont bring me back to GitHub Copilot which I cancelled yesterday. GitHub Copilot had competitive pricing until yesterday when they changed from per-request to one of the most expensive per-token quotas. Seriously, take a…

What application/UI are you using deep seek flash high on? Still copilot or something else

Re: MAI-Code-1-Flash

#177

Post by papascrubs »

GaryBluto wrote: post #174
papascrubs wrote: post #168

Earlier quoted context omitted.

Are you purchasing directly from DeepSeek? Any concerns as far as privacy or data protection?

Using OpenRouter, going to migrate to DeepSeek's official API soon. I'm not using it for anything commercial or for private data so I have no privacy qualms.

Makes sense. Privacy is my only real hang up with DeepSeek. Both of the big SOTA providers have become extremely filtered. Things that I could do one version ago are now getting refusals. Anthropic is almost unusable. ChatGPT is slightly better. Even with a "cyber exception" in place and a vetted account. They are going to force me to take my business elsewhere.

Re: MAI-Code-1-Flash

#180

Post by therealdrag0 »

linuxhansl wrote: post #61
hmokiguess wrote: post #33

Does anyone actually uses these smaller models for coding? If so, how? I usually Opus everything. Is the play to plan/design/architect with a heavier model than delegate structured tasks to these smaller ones? Would appreciate to hear someone's opinion on having done and tested both paths.

I am using Opus 4.x at work, and these "smaller" (20-80bn, 3-4bn active) models at home. Unfortunately there is no comparison, yet (IMHO anyway). With Opus I can work, trust its designs, architecture suggestions, and code changes, even in a complex code base. The smaller models seem to "try". They work for smaller tasks, but for more complex task it's often more work than doing it myself. I wish it were different, an…

Ya when it’s your own token budget on the line the smaller/cheaper models are more attractive.

I’ve used GPT mini quite a bit and it’s decent.

Post reply on HN