Every AI agent ships mountains of context to the model on every request, and most of it is noise. You pay for it, you wait for it, and it doesn't make the output any better.
So we built Paritok — an open-source (Apache 2.0) compression model trained specifically on real coding-agent trajectories. It drops in as middleware between your agent and the LLM API and cuts input tokens by ~74% on a real SWE benchmark, while keeping ~86.5% of solve quality. No rewrites: point Claude Code, Cursor, or any BASE_URL-respecting agent at it and go.
Now we want to see what people build with it, so we're running The Token-Efficiency Hackathon:
$630 in prizes
Build any LLM-powered project — agent, dev tool, RAG app — and use Paritok to make it lean
Free hosted GPU for all hackers, with a live dashboard of your token + cost savings
Fully remote, global, solo or teams
July 20 → Aug 5, 2026
👉 Devpost: https://build-with-paritok.devpost.com/
👉 GitHub: https://github.com/Paritok-official/paritok-4b-v1
👉 Discord: https://discord.gg/SeBJE5Eucp
We're two engineers shipping this on our own budget, so any builder who jumps in genuinely helps — and we're in the Discord all through the hackathon for setup and integration help.