Codeset

Give your AI coding agents the codebase knowledge your team spent years building.

[Codeset @codesetai]

Read more: https://t.co/Xm952x7x6R

[Codeset @codesetai]

We ran the same eval on OpenAI Codex.

Same benchmarks, same methodology as the Claude eval. Results hold.

🧵

[Codeset @codesetai]

Works with Claude Code. Works with OpenAI Codex. Works with all major coding agents.

$5 per repo, one-time.
Code CODESETLAUNCH for a free trial → https://t.co/FgCmiM7PF9

[Codeset @codesetai]

Try it now at https://t.co/FgCmiM7PF9

🎁Free run for the first 50 users with code CODESETLAUNCH

[Codeset @codesetai]

We made Haiku as good as Opus. Without changing the model.

Today we're launching Codeset. Here's what we found 🧵

[Codeset @codesetai]

Results? On our benchmarks, Haiku + Codeset performs better than vanilla Opus at nearly 1/10th the cost!

The next leap in AI coding won't come from smarter models alone. It'll come from models that finally know where they are.

[Codeset @codesetai]

New blogpost is out! Read it at https://t.co/FwvLqwPw2o

[Codeset @codesetai]

Try it now:
🔗 Platform: https://t.co/DUZdsuTgix
📚 Docs: https://t.co/W7C1JR0m4c
💾 Dataset: https://t.co/ZTOvW10obz

$5 free credits, pay-as-you-go, no minimum commitment. We'd love your feedback as we build this out.

[Codeset @codesetai]

We're launching Codeset today 🚀

[Codeset @codesetai]

What's next: more languages, additional task types beyond issue resolution, support for other public datasets, and custom datasets on commission.

We're building this for teams working on the next generation of code agents.

[Codeset @codesetai]

Read more about CWM at https://t.co/m9dASlQLaF

Read more about Codeset at https://t.co/xP7LZC3XQs

[Codeset @codesetai]

🔎 Paper 2 in the Codeset Paper Series:
CWM - An Open-Weights LLM for Research on Code Generation with World Models
by @TacoCohen @adiyossLC @syhw and the Meta FAIR CodeGen Team

[Codeset @codesetai]

We believe that the future of code LLMs lies in understanding computational environments. For this, we need to scale executable data and improve both state observability and verification. Our platform will enable anyone to easily train models that reason about code execution.