There are incredibly powerful new models open source models, and a lot of the coding plans have been tightening and lowering usage. So we are offering free DSV4 flash 0731 and GLM 5.3 Flash for a month on Phoenix Grove API. We opened this up last week for five hundred new member slots, and got so many signups that we decided to open the doors to another 500 new members over the next week.
People are looking for options, and here is one.
Other Cool Stuff:
All of our models are running on 100% US infrastructure, private with zero training on your code or prompts. Use the top open source models without sending your private prompts to a training lab. No complications, no "some models are private, other's aren't". They all are, all the time.
We host 20+ other major models in case you ever want to upgrade (no pressure though). Including the Kimi family, GLM, Qwen, Nemotron and bunch of others. On average our token pricing is 20% lower than market price.
Our higher plans bank up to ten days of usage, so when you aren't using them your usage saves up for later. Usage doesn't go to waste, so you can actually code when you want to.
The intro plan is a free one month trial with the standard cancel anytime, it bills at 3.99 after that. Use it, cancel it, that's fine. Free Flash for a month.
Figured i'd keep this short because we all know the new flash models are the point :)
For the API plan: api.pgsgrove.com
If you want to read more about us as a company, just pgsgrove.com
Also: There's a lot going on in the background with major AI companies right now, we are at a major turning point in the industry.
What's actually happening? This is happening because companies that were purely investment based, now need to answer to their investors. The problem has often been a loss based business model that is finally running dry.
There are several tricks that the major AI coding plans use to extract the most they can from their customers. Here are some examples, and what we are doing differently to put the users first. PGS AI was built with a sustainable business model from the ground up, so we can actually offer great usage rates without tricks.
Wasted usage is part of the AI industry, and they plan on it: Most coding plans bet on you letting usage go to waste. The plan goes: "how do we get people to think our coding plan offers a lot of usage, but then break it up into weeks and rolling windows so no one can ever actually use it all."
Many in app subs and coding plans are glorified training pipelines: This comes along with "how do we harvest this data for training without being too loud about that." Unless the company tells you otherwise, your data could be hopping all over world, being harvested by the individual labs or service companies. Some are better than others, but many of these companies rely on users just not noticing or caring that their data is being used for training. Data sales and marketing telemetry sales happen. This means that your private info, your personal life, and anything else you send through the system could become part of a training corpus for the next AI, or a marketing data set for a large company.
Privacy and ease of use should be available for everyone. It's too often a trade off, and we are hoping to see that change.
[link] [comments]