Perplexity open-sources pplx-embed-v2 embeddings, with 9B model hitting 92.4% on multimodal retrieval
antoine_chaffin · x · 2026-10-08
Perplexity's team has released pplx-embed-v2, a second-generation collection of embedding models now live and open on Hugging Face.
- Two sizes: 9B (late-interaction) and 0.6B, plus a pplx-embed-v2-context-9b-preview
- Results: the 9B model sets a new SOTA among retrievers at 92.4% accuracy on multimodal retrieval, while also offering frontier-level text retrieval
- Asymmetric design: shared embedding space enables cross-model retrieval and flexible production deployments
The author invites the community to try the models and share feedback.
Related event: Perplexity Open-Sources Multimodal Embedding Models pplx-embed-v2-late(38 posts)→
More from Models
- ChatGPT on Browser Shows 'Capabilities Reduced' Warning — A New Kind of Rate Limit? — jasondeanlee · 2026-10-08
- ChatGPT Work mode vs Codex: same quota, far more tasks done per 5-hour window — sasik520 · 2026-10-08
- Epoch's InnovationEval: AI agents still far from producing real research innovations — Afinetheorem · 2026-10-08
- 113 decision models in 3 weeks: 70 built on Qwen, sub-cent per call — jonathanmalkin · 2026-10-08
- User reports Haiku 5.5 is a major workflow upgrade in screenshot post — Sorcerer12345 · 2026-10-08
- OpenRouter launches Decision Model Rankings, with typesafeai leading all categories — gaganghotra_ · 2026-10-08