~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.
Python academiaanthropicarxivbravedeep-researchencryptionhome-automationhomeserverlocallocal-deep-researchlocal-llmmistralollamaopenaipubmedresearchresearch-toolretrieval-augmented-generationsearxngself-hosted
test(research): pin lower bound on the with-query log-preview probe (#7368)
Ziyang Guobf59fbc80b98
diff --git a/changelog.d/7109.misc.md b/changelog.d/7109.misc.mdnew file mode 100644index 0000000000..7f8a60bc29--- /dev/null+++ b/changelog.d/7109.misc.md@@ -0,0 +1 @@+Test hardening: the with-query log-preview probe in `test_research_model_log_preview_bound.py` now also asserts `CountingValue.reprs >= 30` alongside the existing `< 200` ceiling. The previous upper-bound-only pin would pass vacuously if the query-length gate ever moved above the log lines the probe targets; the floor sits below the measured count of 40 so it fails the probe when the bounded previews stop running rather than passing vacuously (#7109).