Wait how do I apply?

Cc @coldniko

You’re missing out by not using Strata You could hit 100+tps on Qwen3.8 Flash Next and 10+tps on DeepSeek V4.1 Flash 1017 (estimates) Flash Next is much better than 27b as well

No. Qwen3.8 Flash Next will, because of Srata Seriously, hats off to @coldniko, 70tps+ on hardware under $3k with models less than 6 months behind the frontier is unreal

Imagine strata with DeepSeek on here 🤤

What models are supported? M1?

Have you looked at Eagle-4?

It’s because large language models are trained on the average of all ideas. When prompted correctly, especially with Claude models, you can get them to come up with new things, but it’s difficult The AI he’s describing is something that is not an LLM, but we already have all of the technology to run it

@coldniko Seems like a great size for Strata 👀

TestFlight please 🙏

@AmpCode orbs. Zero contest.

Being bullish on Mistral is a recession indicator

Codemode

Last I checked 135iq is more than 115 unless I’m blind

🙏 I am

Try project maya, you should be able to get 50+tps

You planning to maintain this?

Lisan and Theo belong higher

If it’s faster and cheaper per token I don’t really mind the token consumption, it still ends up faster and cheaper overall Even if the official provider is slow there are others (baseten) with 5x the speed of sol per token, it balances out And as you mentioned, swarms

What skills are you testing? Tasteskill v2?