Codebuff solves at least 40% of issues on SWE-Bench by March 31, 2025
3
1kṀ1261Mar 31
16%
chance
1D
1W
1M
ALL
(This market is AI-generated but I read it and it seems right)
This market predicts whether Codebuff will achieve a 40% success rate on the SWE-Bench dataset, which is a benchmark of human-selected program issues.
Resolution will be based on official results published on the SWE-Bench dataset or Codebuff project's official channels.
References:
This question is managed and resolved by Manifold.
Get
1,000 to start trading!
1,000Sort by:
Nice. We can do it!
I assume you mean the full SWE bench. We're more likely to work on the Lite or Verified subset.
Related questions
Related questions
Will any model get above human level on the Simple Bench benchmark before September 1st, 2025.
45% chance
AI resolves at least X% on SWE-bench without any assistance, by 2028?
AI resolves at least X% on SWE-bench WITH assistance, by 2028?
What will be the best performance on SWE-bench Verified by December 31st 2025?
When will SWE-bench be solved?
Will an autonomous agent resolve 90% of tasks on SWE-bench by 2026?
63% chance
What will be the highest score achieved on SWE-Bench Verified in 2025?
What will be the best score on Cybench by December 31st 2025?
Will Alphaproof achieve >30% performance on the FrontierMath benchmark before 2026?
22% chance
Top SWE-Bench Verified score in 2025?
-