PTEB: leaderboard
Persian Text Embedding Benchmark leaderboard evaluating embedding models across Persian pair classification, retrieval, reranking, intent, scenario, and sentiment tasks.
Metric: Mean Score (%). Source: huggingface.co. 34 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | F2LLM-v2-4B | 77.49 |
| 2 | embeddinggemma-300m | 76.87 |
| 3 | F2LLM-v2-1.7B | 76.52 |
| 4 | multilingual-e5-large-instruct | 76.13 |
| 5 | jina-embeddings-v5-text-nano | 76.1 |
| 6 | jina-embeddings-v5-text-small | 76.07 |
| 7 | Qwen3-Embedding-4B | 75.37 |
| 8 | bge-m3 | 75.07 |
| 9 | F2LLM-v2-0.6B | 74.5 |
| 10 | jina-embeddings-v3 | 73.63 |
Interactive version: theaggregate.ai/benchmark?slug=pteb · How It Works · Data refreshed daily, snapshot 2026-09-05.