Felipe Domingues

awesome things in • portfolio • dad • tiny/small LLM Researcher

[Felipe Domingues @_domfelipe]

@Da7_Tech I agree, and cuz of it I'm developing my own harness, based on Grok Build with other traces, imo is the best option today. But we have good harnesses as well, like droid, pi, and the official ones.

[Felipe Domingues @_domfelipe]

@Da7_Tech Oh, i saw it today, the medium quota drain tokens so fast, like 2h and it ends

[Felipe Domingues @_domfelipe]

@osanseviero NEW GEMMA's? 🎉🎉

[Felipe Domingues @_domfelipe]

@MiaAI_lab will be our new base model for fine tuning! Awesome!

[Felipe Domingues @_domfelipe]

@bindureddy Nobody is saying that's a frontier model, but it is a mega honest model, good performance with a very low price

[Felipe Domingues @_domfelipe]

@StatsWire 10k tasks a day is insane, more than the extremely low cost of it.

[Felipe Domingues @_domfelipe]

@V1adXBT Hey guys! Let's talk about AI?

[Felipe Domingues @_domfelipe]

@ardent__dev @ChatGPT for work and personal, thats a proud of @thsottiaux

[Felipe Domingues @_domfelipe]

@adonis_singh The best is to use the frontier model that you have access to, even Kimi K3 or Gemini 3.6 Flash (that's a good model to orchestrate in my personal tests)

[Felipe Domingues @_domfelipe]

@GoogleAIStudio https://t.co/b7wmM0gb5W
My own benchmark platform (still in dev mode yet)

[Felipe Domingues @_domfelipe]

The new qwen 3.8-27B can be a base frontier model for fine-tuning small models, like the 3.6 is nowadays. Can't wait to see it!

[Qwen @Alibaba_Qwen]

📢Meet Qwen3.8-Max — our most capable model to date.

Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉

Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters:

- Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub：https://t.co/iVHZWQoeSo
- Real work, real results: Production-quality deliverables across hundreds of professions.
- Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy.
- Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction.

💰Pricing:
Input: $2.0 / M tokens
Output: $6.0 / M tokens
Implicit Caching: $0.25 / M tokens

Start building with Qwen3.8-Max! 🚀

📖 Blog: https://t.co/iwjmQxLBof
✅ Qwen Studio: https://t.co/4V2pFvDovG
⚡ API: https://t.co/gAGqaLQGbN

[Felipe Domingues @_domfelipe]

@Alibaba_Qwen Already available in the token plan? ;)

[Felipe Domingues @_domfelipe]

I think the battle on performance/cheaper price in AI could be more interesting, if @AnthropicAI release a model like GPT 5.6 Luna and now DSV4-flash. Perhaps a new haiku?

Imo, by now Luna is a cheat mode. Thanks @OpenAIDevs

[Felipe Domingues @_domfelipe]

@Aditya_181105 The best is and always be that the fit in your budget

[Felipe Domingues @_domfelipe]

@Nozelcode https://t.co/18suxGHzyG

[Felipe Domingues @_domfelipe]

@NanouuSymeon Everyday is ChatGPT/Codex. Don't you @thsottiaux

[Felipe Domingues @_domfelipe]

@blueemi99 Don't remember them dude...

[Felipe Domingues @_domfelipe]

@thsottiaux @thsottiaux Sry the topic ress, But we could have a tab called "orchestration" to make it easier to distribute tasks between @OpenAI models (or other external models, why not?) And improve token management. The layout could be similar to @evertjr. Hey, @sama , notice me!