# Open-source lemmatizing analyzer for ES 9.x: +18.8% F1@10 vs standard on MIRACL English (reproducible benchmark) **URL:** \ **Category:** Community Ecosystem\ **Created:** [November 10, 2026, 10:23pm UTC](https://discuss.elastic.co/t/open-source-lemmatizing-analyzer-for-es-9-x-18-8-f1-10-vs-standard-on-miracl-english-reproducible-benchmark/391510 "2026-11-10T22:23:26Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![tonyj](https://avatars.discourse-cdn.com/v4/letter/t/e9c0ed/32.png) [@tonyj](https://discuss.elastic.co/u/tonyj)\ **Post date:** [November 10, 2026, 10:23pm UTC](https://discuss.elastic.co/t/open-source-lemmatizing-analyzer-for-es-9-x-18-8-f1-10-vs-standard-on-miracl-english-reproducible-benchmark/391510/1 "2026-11-10T22:23:26Z")
Hi all, We measured how much BM25 relevance depends on the analyzer alone, and wanted to share the results and the plugin with the community. Setup: Elasticsearch 9.5.4, plain BM25, MIRACL English (32.9M passages, 799 queries), official trec\_eval. - vs standard: F1@10 +18.8%, nDCG@10 +15.7%, Recall@100 +14.8% - vs built-in english (Porter): +7.6% to +8.5% on every top-10 metric The main gains come from dictionary lemmas instead of stemming: Porter can't map sold→sell or built→build, and it over-merges majority→major and United→unit. Limits: one dataset, English only, lexical only. Disclosure: I work at Bitext, which built the analyzer. It's Apache 2.0 and installs as an ES 9.x stable plugin (also OpenSearch 3.4+ and Solr 10). Benchmark (fully reproducible): [GitHub - bitext/bitext-search-benchmarks: Reproducible MIRACL lexical-search benchmarks to measure the benefits that the Bitext English Analyzer adds to Elasticsearch · GitHub](https://github.com/bitext/bitext-search-benchmarks) Feedback is welcome, especially cases where it does worse than the english analyzer on your own data.