beir
tool judged 10% likely a named technique or architecture
See how it has spread across media tiers
Mentions over time
Table view
| Week | Articles | Sources |
|---|---|---|
| 2024-W22 | 1 | 1 |
| 2024-W33 | 1 | 1 |
| 2026-W24 | 1 | 1 |
| 2026-W25 | 1 | 1 |
| 2026-W32 | 1 | 1 |
| 2026-W36 | 1 | 1 |
| 2026-W39 | 1 | 1 |
In their words
We’re sharing Constella as a research preview, with results across 15 BEIR datasets.
In our production evaluations on the BEIR benchmark, adding any of these rerankers on top of dense retrieval lifts NDCG@10 by 8–14 points, a bigger jump than switching embedding models.
This tool supports 22 benchmarks in total, further including benchmarks such as BEIR, LoTTe, EnronQA, and FreshStack if you would like to extend these comparisons.
First, retrieval accuracy, measured by NDCG@10 or recall@10 on benchmarks like MTEB v2 and BEIR.
Leading Rerankers in 2026# | Reranker | Context Window | Key Feature | |---|---|---| | Cohere Rerank 3.5 | 4K tokens | Production baseline, handles tables/JSON/code | | Voyage rerank-2.5 | 32K tokens | Instruction-following — steer relevance with natural language | | bge-reranker-v2 | 8K tokens | Open-source, self-hostable, strong on BEIR benchmarks | Voyage rerank-2.5's instruction-following capa
We use Milvus and snowflake-arctic-embed-m model, which achieves state-of-the-art performance on the MTEB/BEIR leaderboard for each of their size variants.
We use Milvus and snowflake-arctic-embed-m model, which achieves state-of-the-art performance on the MTEB/BEIR leaderboard for each of their size variants.