Ofir Press @OfirPress · 01 Sep 2026
Releasing SWE-bench Multimodal v2.0 today 480 tasks where a coding agent must interpret visual assets like screenshots, diagrams, recordings to diagnose and fix a bug in a repository. https://t.co/tEvvRFzwCr
22 786Views
238Likes
35Reposts
15Replies
2Quotes
0Bookmarks
Is that a lot?
7.11×vs this author's median3 207 views is typical
82Percentile for this authorof 11 recent posts
6.33×vs 10K–100K median3 598 views is typical
116.91%Reachviews ÷ followers
1.27%Engagement rateof viewers reacted