architect by day, tinkerer by night. ≥ https://teapawt.net/t.co/kmsJMKOyJH ≥ https://teapawt.net/t.co/v61NkbBarz ≥ https://teapawt.net/t.co/Zd5g7Oryh5
Spain
Joined July 2022
- Tweets836
- Following268
- Followers231
- Likes190
Kimi K3.1 Pro joins FarmerBench 9th of 14: a mean of 48.1 over three seeds, between Claude Sonnet 5.5 (55.1) and DeepSeek V4.1 Flash (46.8).
FarmerBench gives a model six game-hours to build an economy in The Farmer Was Replaced, by writing code.
static analysis + proper tests (not GPT's test for test sake) get you a good part of the way there
i feel like theres a gap somewhere for a judgement agent like we see in evaluation frameworks that will help bridge a gap traditional tooling can't
I wonder if this will overtake Argon on FarmerBench :O
farmerbench.dev/models/gemin…
Gemini 4 Argon is currently #1 on FarmerBench. I wonder where this will place..
farmerbench.dev/models/gemin…
Gemini 4 Argon is currently the highest rated model on FarmerBench and I think had a really unfair reception here on X.
Super interested to see how Gemini 4 Carbon looks.
Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago.
Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned.
working on farmerbench.dev right now, enjoying playing with harnesses and benchmarking models!
building model benchmarks and other random stuff in public!
lets connect
farmerbench.dev
building model benchmarks and playing with cool AI stuff
what you working on?
farmerbench.dev
working on farmerbench.dev and playing with models/harnesses!
lets connect!
How it's made: Three.js + Rapier physics, baked at 240Hz and rendered frame by frame in headless Chrome. Each coin is a follower's profile pic. The metal depends on your own follower count, and the wear depends on how long you've followed. About 5 minutes to render; considerably longer to stop fiddling with it.