Vic Klein
@its_hapenin
Followers
93
Following
82
Media
2
Statuses
48
research and things @symbolica
London
Joined December 2025
I’ve been working on ARC-AGI-2 and the best part of this for me was seeing how different models behaved, how they worked it out and when they didn’t. You can see the logs here. Thanks to @fchollet for the challenge.
symbolica.ai
Exploiting the reasoning capabilities of code mode agents and RLMs with the Agentica framework.
10
4
24
Introducing Mixedbread Wholembed v3, our new SOTA retrieval model across all modalities and 100+ languages. Wholembed v3 brings best-in-class search to text, audio, images, PDFs, videos... You can now get the best retrieval performance on your data, no matter its format.
35
120
945
We've released the code used to solve all 3 publicly available ARC-AGI-3 games. https://t.co/hFogxNbIRr
github.com
Contribute to symbolica-ai/ARC-AGI-3-Agents development by creating an account on GitHub.
6
36
356
Join us this Thursday, 26th in London. We’ll be speaking about ARC-AGI-2 and ARC-AGI-3 at the London AI nerd meetup: what’s new, what’s changing, and what it means for agents. @its_hapenin @FireworksAI_HQ
https://t.co/y50r2z5s7k
2
3
22
this happened while we ate lunch
1
2
10
👀👀
Really cool to see teams already working on ARC-AGI-3 We'll have public replay links for every online run
0
0
8
If you’re interested :)
Vic (@its_hapenin) will chat ARC-AGI-2 results at Engineering Night London - what worked, what didn’t, and what surprised us. With Agentica: • 85.28% with Opus 4.6 (120k) High • +10pp GPT-5.2 (XHigh) vs CoT • +20pp Opus 4.5 vs CoT See you there. Register:
2
0
6
Funnily enough I tried to dabble with ARC AGI before and with very little success… Super cool to well designed RLMs achieving SOTA :)
We set a new ARC-AGI-2 SotA: 85.28% using an Agentica agent (~350 lines) that writes and runs code.
5
13
218
1
0
7
i wrote a thing with @charlielidbury, because agents want an execution environment https://t.co/CkaiaQqf7D (we built it too)
3
4
12