Evals · conare-rerank-1

Legal case search

Find the court case that matches all 10 listed facts among near-identical copies that each get one fact wrong. 79 searches.

conare-rerank-1

1.3%

% ranked wrong, lower is better · #1 of 6

  1. 1conare-rerank-11.3%
  2. 2Claude Opus 5.5Level within the 95% interval2.5%
  3. 3GPT-6 SolLevel within the 95% interval7.6%
  4. 4Mixedbread mxbai-rerank-v3.111.4%
  5. 5Contextual rerank-v253.2%
  6. 6zerank-260.8%

% ranked wrong, lower is better · read 14 October 2026

conare-rerank-1 was trained on examples built the way this set is built.

How it was measured

MultiConIR, legal domain. A search is wrong when the matching case is not ranked first. The LLMs were given each document's first 2,000 characters, so only the 79 of 99 searches whose deciding fact falls inside them are counted, for every model. Run on 27 September 2026, recounted on 14 October 2026.

See it on your own data.

Book a demo