MTEB
mteb.org
News from MTEB: new features, models, results, datasets and benchmarks. Every number is recomputed from public results — check them yourself at leaderboard.mteb.org. Posts are auto-drafted from GitHub activity, but curated by maintainers.
👤 Joined on 03 August 2026
38
Followers
5
Following
27
Total Posts
📜 Kullanıcı Geçmişi
- 03 Ağu 2026, 14:05
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@mteb-leaderboard.bsky.social - 03 Ağu 2026, 14:46
✏️ Kullanıcı adı değişti:
@mteb-leaderboard.bsky.social→@mteb.org
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
1/ When we built MAEB, we cut audio at 30 seconds per clip to keep evaluation affordable. Many models support much longer audio, and at least 16 of our audio tasks have longer clips (up to 21 minutes), so those models were only scored on the start of each recording.
Show Post2/ From mteb 2.24.4, each audio and video model uses the limits its authors set instead: its own maximum clip length, sampling rate and number of video frames. This is what we already do for text, where each model reads up to its own maximum input length.
Show Post3/ Because scores change, we will rerun the audio and video models on the leaderboard with the new settings. If you submit new audio or video results, please use mteb 2.24.4 or later. Work by AdnanElAssadi.
Show Post
📈 Google's EmbeddingGemma2 is looking to be a huge success! After only a week, it has surpasses EG1's likes to become the 4th most liked Google model ever, as well as the 3rd most liked Sentence Transformers model ever. Congratulations to the team for their work! Links in 🧵
Show Post
The new bekko embedder based on a pruned ModernBERT is neat but they market Matryoshka while neglecting quantization; once again quantizing the vectors is far more effective
Show Post