AIEvals

Benchmarks

AA-Omniscience (public split)

Factual knowledge and hallucination on the publicly available AA-Omniscience questions.

Factual knowledgeArtificial Analysis, evaluated and reported by Reflection AIVendor-reported onlyReported by Reflection AI

How it is run

Reflection AI explicitly labels this an index on the public split only, calculated from its own results. Five model scores are reported.

What it cannot tell you

The unit is an index score, not a percentage or the AA Intelligence Index. The public-only result is separate from AA’s independently published full evaluation.

Why there is no independent table

No independent suite in this dataset runs AA-Omniscience (public split). The numbers below are the ones model publishers report about their own models, which is a different kind of evidence: each publisher chooses its own harness and effort level, so a row is readable on its own and not against the other rows.

Read this with

Other versions of the same evaluation. Versions are not comparable: the task set changes between them.

  • AA-OmniscienceSame family. Factual recall and hallucination across economically relevant domains.

As reported by each publisher

Kept separate per publisher, because the settings are not shared.

Reflection AI

Introducing Beam: model capability comparisons · 5 October 2026

RankModelScore (score)
3Beam13
2Inkling14.2
4Nemotron 3 Ultra8.6
1GLM 5.322.6
5DeepSeek V4.1 Flash6.6

Ranks compare reported scores in the same published row; ties share a rank.

AA Omniscience (index) - public split only (our results). Reflection AI’s public-only index, not AA’s independently evaluated full score or the Intelligence Index.

Reflection AI’s four capability tables, read 6 October 2026. All 21 rows and eight model columns are transcribed; NR means not reported and remains blank. The source’s Qwen 3.8 Max column uses this catalogue’s Qwen3.8 Max spelling. Beam results are publisher reports; Reflection AI cites Artificial Analysis and Datacurve for comparison-model evaluations. The announcement does not specify a common harness, effort or trial count for every row. Ranks are derived within each published row from reported values only, with ties sharing a rank; they do not establish an independent or overall model ranking. AutomationBench uses the public split; AA-Omniscience is an index on the public split only, calculated by Reflection AI. BrowseComp and DeepSearchQA use context management. AA-LCR has no stated version. These settings remain separate from other source fields.

reflection.ai/blog/introducing-beam

Sources

Every page used for this benchmark, read on 17 October 2026.