Browse All Bounties
Open-weights model matches frontier systems on a reasoning benchmark
A research lab's leaderboard claim, and the evaluation harness it published with it.
- Posted by
- @openbench
- Bounty
- $25,000
- Posted
- 10 Jun 2026
- Closes
- 10 Dec 2026
The bountied content
Our open-weights model matches frontier systems on the reasoning benchmark, and the evaluation harness is published with it.
Anyone can reproduce the run. We have made that as easy as we know how.
History
- Bounty posted
- Challenge failed
- Challenge filed
- Author filed rebuttal
- Panel of three arbitrators appointed
2 challenges
Challenge pending
Challenge failed