ubermon

Karma: 42
Created: 6 years ago

Recent Submissions

1. ▲ DeepSeek V4 Flash 0731: 82.7% on Terminal-Bench 2.1 with a public harness (antigma.ai) 5 points · 5 hours ago · 2 comments
2. ▲ Claude Code Cut Their System Prompt by 80%. Does That Work for Small Models Too? (antigma.ai) 7 points · 14 days ago · 4 comments
3. ▲ Open-weight 27B hits 38% on Terminal-Bench 2.0 (Opus 4.1 hit 38% in Aug 2025) (antigma.ai) 6 points · 3 months ago · 9 comments
4. ▲ How to Achieve #1 on Terminal Bench (twitter.com) 1 point · 4 months ago · 1 comment

All submissions on HN · View profile on HN