pplx-embed-v2-late 0.6B/9B, EmbeddingGemma 2, Cohere Embed 5 Pro/Fast, and pplx-embed-v2-context-9B Preview.
| Rank | Release Date | Model & Lab | Category | Architecture | Score | Action |
|---|---|---|---|---|---|---|
| #4 | Oct 7, 2026 |
pplx-embed-v2-late 0.6B
Perplexity
|
Agent Memory, Orchestration & Infra | Late-Interaction Contextual Embedding Transformer Β· 0.6B Params | 100.7 | Try β |
| #5 | Oct 7, 2026 |
pplx-embed-v2-late 9B
Perplexity
|
Agent Memory, Orchestration & Infra | Dense Late-Chunk Retrieval Representation Transformer Β· 9B Params | 101.6 | Try β |
| #7 | Oct 6, 2026 |
EmbeddingGemma 2
Google
|
Agent Memory, Orchestration & Infra | Dual-Encoder Open Vector Representation Model Β· Gemma 2 Backbone | 102.8 | Try β |
| #22 | Sep 30, 2026 |
Embed 5 Pro
Cohere
|
Agent Memory, Orchestration & Infra | Dense Semantic Vector Representation Model for Enterprise Search | 73.8 | Try β |
| #23 | Sep 30, 2026 |
Embed 5 Fast
Cohere
|
Agent Memory, Orchestration & Infra | Low-Latency High-Throughput Vector Embeddings for Real-Time RAG | 73.1 | Try β |
| #24 | Sep 30, 2026 |
pplx-embed-v2-context-9B Preview
Perplexity
|
Agent Memory, Orchestration & Infra | Long-Context 64k Vector Representation Preview Transformer Β· 9B Params | 69.5 | Try β |