# Document Reranking on BEIR (NDCG@10) Source: [Document Reranking on BEIR (NDCG@10)]() Collection index: [llms.txt]() Data updated: 2026-03-05T08:16:38.899Z Evaluation of Document Reranking performance across the BEIR benchmark suite components (including Covid, NFCorpus, Touche, DBPedia, SciFact, Signal, News, Robust), measured using NDCG@10 for each sub-dataset. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Document Reranking | | Dataset (including any stated split) | BEIR | | Metric | Average NDCG@10 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Average NDCG@10](): Higher is better; unit: PERCENT. - [NDCG@10 (Covid)](): Higher is better; unit: PERCENT. - [NDCG@10 (NFCorpus)](): Higher is better; unit: PERCENT. - [NDCG@10 (Touche)](): Higher is better; unit: PERCENT. - [NDCG@10 (DBPedia)](): Higher is better; unit: PERCENT. - [NDCG@10 (SciFact)](): Higher is better; unit: PERCENT. - [NDCG@10 (Signal)](): Higher is better; unit: PERCENT. - [NDCG@10 (News)](): Higher is better; unit: PERCENT. - [NDCG@10 (Robust)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Average NDCG@10 | Result date | Report paper | | --- | --- | --- | --- | | DeBERTa-v3 | 55.4 | Not recorded | 68fa77ee1e209131cea76195 | | DiffuRank Perm-Assign | 55.21 | Not recorded | 6992a4b39ebc499c97406d75 | | DiffuRank Perm-Samp | 55.1 | Not recorded | 6992a4b39ebc499c97406d75 | | OSCAR-10-Layers | 54.3 | Not recorded | 68fa77ee1e209131cea76195 | | Llama-3.1-8B Listwise | 54.28 | Not recorded | 6992a4b39ebc499c97406d75 | | DiffuRank Logits-List | 54.09 | Not recorded | 6992a4b39ebc499c97406d75 | | Qwen3-8B Pointwise | 53.92 | Not recorded | 6992a4b39ebc499c97406d75 | | Qwen3-8B Listwise | 53.39 | Not recorded | 6992a4b39ebc499c97406d75 | | OSCAR-llama | 52.8 | Not recorded | 68fa77ee1e209131cea76195 | | DiffuRank Pointwise | 52.68 | Not recorded | 6992a4b39ebc499c97406d75 | | OSCAR-8-Layers | 52.5 | Not recorded | 68fa77ee1e209131cea76195 | | LLaDA-1.5-8B Listwise | 52.48 | Not recorded | 6992a4b39ebc499c97406d75 | | Llama-3.1-8B Pointwise | 51.68 | Not recorded | 6992a4b39ebc499c97406d75 | | RankZephyr-7B Listwise | 51.15 | Not recorded | 6992a4b39ebc499c97406d75 | | monoT5 (220M) | 46.44 | Not recorded | 6992a4b39ebc499c97406d75 | | monoBERT (340M) | 45.55 | Not recorded | 6992a4b39ebc499c97406d75 | | BM25 | 43.43 | Not recorded | 6992a4b39ebc499c97406d75 | ### DeBERTa-v3 Average NDCG@10: 55.4. Result ID: 69a93bd46dbb7075b9528dea. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"role":"teacher"}. This record does not assert independent reproduction. ### DiffuRank Perm-Assign Average NDCG@10: 55.21. Result ID: 699a57de5698665ec4df899b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaDA-1.5-8B","Strategy":"Perm-Assign","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### DiffuRank Perm-Samp Average NDCG@10: 55.1. Result ID: 699a57de5698665ec4df899a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaDA-1.5-8B","Strategy":"Perm-Samp","Sampling Steps (K)":2,"Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### OSCAR-10-Layers Average NDCG@10: 54.3. Result ID: 69a93bd46dbb7075b9528de9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Decoder":"Mistral-7B"}. This record does not assert independent reproduction. ### Llama-3.1-8B Listwise Average NDCG@10: 54.28. Result ID: 699a57de5698665ec4df8994. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Strategy":"Listwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### DiffuRank Logits-List Average NDCG@10: 54.09. Result ID: 699a57de5698665ec4df8999. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaDA-1.5-8B","Strategy":"Logits-List","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### Qwen3-8B Pointwise Average NDCG@10: 53.92. Result ID: 699a57de5698665ec4df8995. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Strategy":"Pointwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### Qwen3-8B Listwise Average NDCG@10: 53.39. Result ID: 699a57de5698665ec4df8996. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Strategy":"Listwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### OSCAR-llama Average NDCG@10: 52.8. Result ID: 69a93bd46dbb7075b9528de7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Decoder":"Mistral-7B"}. This record does not assert independent reproduction. ### DiffuRank Pointwise Average NDCG@10: 52.68. Result ID: 699a57de5698665ec4df8998. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaDA-1.5-8B","Strategy":"Pointwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### OSCAR-8-Layers Average NDCG@10: 52.5. Result ID: 69a93bd46dbb7075b9528de8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Decoder":"Mistral-7B"}. This record does not assert independent reproduction. ### LLaDA-1.5-8B Listwise Average NDCG@10: 52.48. Result ID: 699a57de5698665ec4df8997. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaDA-1.5-8B","Strategy":"Listwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### Llama-3.1-8B Pointwise Average NDCG@10: 51.68. Result ID: 699a57de5698665ec4df8993. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Strategy":"Pointwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### RankZephyr-7B Listwise Average NDCG@10: 51.15. Result ID: 699a57de5698665ec4df8992. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Zephyr-7B","Strategy":"Listwise","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### monoT5 (220M) Average NDCG@10: 46.44. Result ID: 699a57de5698665ec4df8991. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"T5","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### monoBERT (340M) Average NDCG@10: 45.55. Result ID: 699a57de5698665ec4df8990. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"BERT","Fine-tuning dataset":"MS MARCO"}. This record does not assert independent reproduction. ### BM25 Average NDCG@10: 43.43. Result ID: 699a57de5698665ec4df898f. [Reporting paper]() Recorded evaluation configuration: {"Type":"Initial Retrieval"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.