Stømb

[Stømb @Stoemb]

Excited to introduce a project I've been working on: https://t.co/PADqpLuIcd! 🚀

It's a free tool for building, visualizing, and running powerful AI reasoning workflows.

Go beyond single prompts and start orchestrating multiple AI models to solve complex problems.

[Stømb @Stoemb]

@soccertorey di maria

[Stømb @Stoemb]

@rezoundous all weekly credits used we need to wait a week

[Stømb @Stoemb]

@tanishqkumar07 @tri_dao @avnermay Where is SSSD?

[Patrick Loeber @patloeber]

:) :)

[Stømb @Stoemb]

Who else is confused by the @xai releases is Grok-4.1 Fast the same model as Grok-4.1?

[Stømb @Stoemb]

Impressive scores for Grok 4.1 thinking and none on LMArena

[Arena.ai @arena]

🚨Text Leaderboard Update

@xAI’s Grok 4.1 (thinking) and Grok 4.1 have scaled new heights in the most competitive Text Arena:
🔹Grok 4.1 (thinking) lands at #1 with a score of 1483
🔹Grok 4.1 follows at #2 with a score of 1465

On the Arena Expert leaderboard:
🔸Grok 4.1 (thinking) also ranks at #1 with a score of 1510
🔸Grok 4.1 ranks at #19 with score of 1437

This is a 40+ point improvement since Grok 4 fast, which landed in the Arena just two months prior.

Congrats to the @xAI team for this incredible milestone! 👏

[Stømb @Stoemb]

@VraserX it is just a psyop

[Stømb @Stoemb]

yes totally agree in europe

[sui ☄️ @birdabo]

engineering is the most important field in humanity.

[Stømb @Stoemb]

@bindureddy in the end it will be only one model

[Stømb @Stoemb]

@VraserX drinking beer at 10 am on a monday

[Stømb @Stoemb]

@JosephJacks_ transformers aren’t the problem in Yann‘s eyes only the autoregressive generation

[Stømb @Stoemb]

@tunguz now i am excited for Gemini 3

[Stømb @Stoemb]

Build in public: AI Strava assistant.

→ Pulls my run data via API
→ Analyzes performance trends
→ Recommends personalized workouts
Already giving better recs than expected.

Should I open-source it?

[Stømb @Stoemb]

@haider1 When comparing the results of GPT 4o and GPT 5 on LMArena i can understand why people miss GPT 4o especially in a more social use case. GPT 4o‘s answers have a lot more warmth. But this could this be fixed for GPT 5 with updates

[Stømb @Stoemb]

@kimmonismus Systems like AlphaEvolve may be a direction which could lead to a soon-ish AGI. I think it’s mostly a engineering problem. Make LLMs more independent, give them the right tools to explore ideas

[Stømb @Stoemb]

@haider1 the goalpost is getting moved again

[Stømb @Stoemb]

@scaling01 more confusing then before. Naming is better but still confusing

[Stømb @Stoemb]

@kimmonismus I consistently triggered reasoning by adding "Think for longer." Without it, the output differs significantly.

[Stømb @Stoemb]

@SmokeAwayyy It was a Wednesday, and I had training.

[Stømb @Stoemb]

Create your own Pro Mode reasoning flow with the new 'gpt-oss' models on https://t.co/XqJdUqwNvm. Just get a openrouter api key, plug it in, create your reasoning flow, select the models and go