HN
Hacker News
Top
New
Best
Ask
Show
Jobs
Comment by byyoung3 | Hacker News Reader
Full thread
byyoung3
·
looks like code llama 70B outperforms on humaneval I believe
View on HN
SparkyMcUnicorn
·
Looks like even deepseek-coder 1.3b benchmarks higher than CodeLlama 70b.
https://evalplus.github.io/leaderboard.html
Reply on news.ycombinator.com