Nishi Deep Research

A sovereign, Co-Scientist-class async multi-agent research engine — deterministic, integer/no-float, bit-exact reproducible on any machine. One call runs the whole loop, over a real on-disk corpus, with a semantic judge.

11 tools liveone-call end-to-endsemantic judgematurity L2no external deps

Grounded against verified 2026 SOTA — Google Co-Scientist (Nature 655:487), DRBench (ServiceNow, ICLR 2026) and DEER (LG AI), Structured-Test-Time-Scaling and FutureWeaver.

Measured against real DRBench groundtruth — 100 tasks

ServiceNow publishes DRBench's task data. We fetched all 102 published task directories over sovereign TLS and scored 100 of them, using ServiceNow's own labels (qa_type=insight vs qa_type=distractor) against each task's real research question. Nothing was self-authored.

The two judges use different scales, so they are compared at matched precision: thresholding just above each judge's own highest-scoring distractor yields zero false positives — 100% distractor avoidance for both. At that identical, perfect precision:

JudgeMean insight-recallDistractor avoidance
lexical254‰1000‰
semantic306‰1000‰

Per task, the semantic judge is better on 32, the lexical judge on 17, and they tie on 51 — roughly a 2:1 win ratio, for a mean gain of +52‰ (about +20% relative).

Correction. An earlier version of this page reported a single-task result (DR0002: 2 of 7 vs 4 of 7) and described the semantic judge as doubling insight recall. That task was not representative, and it used the opposite scoring direction. Across the full 100-task set the real effect is about +20%, not +100%. The figures above supersede it.

Scope, stated plainly: this is not a DRBench leaderboard score and is not comparable to published agent numbers. It measures a judge discriminating labelled answer texts — not an agent performing research and writing a report over the enterprise environment. The enterprise half requires the Docker stack, which was not run. No source documents were fetched (publisher sites bot-block), so no source-grounded recall was computed — and source text was deliberately not synthesised from the groundtruth answers, because scoring groundtruth against itself would be rigged.

Live capabilities (each gate-proven + verified via a live tools/call)

ToolRoleGate
nx_dr_semjudgeSemantic judge — PPMI late-interaction over the real 102,318-word model9/9
nx_dr_corpusCorpus mode — real files on disk; separates corpus gaps from budget gaps; fail-loud on unreadable8/8
nx_dr_chainCapstone — ONE call runs judge → conduct → score, reporting the measured cost of the budget9/9
nx_dr_runLexical judge — tokenizer + containment9/9
nx_dr_conductorAsync conductor — verify → Elo tournament → opportunity-cost dispatch10/10
nx_dr_eloTournament-of-Ideas ranker — deterministic Elo9/9
nx_dr_ocmOpportunity-cost ranker — (Value×Momentum)÷Cost + budget selection9/9
nx_dr_verifyCitation-entailment loop — drops unsupported claims9/9
nx_dr_refuteAdversarial refutation — decorrelation discount9/9
nx_dr_drbenchDRBench+DEER scorer — insight-recall, avoidance, factuality, completeness, evidence-validity10/10
nx_dr_routeWeb-vs-private router — private-insufficient queries never dead-end9/9

The loop, in one call

nx_dr_corpus / nx_dr_chain ─┐
   ├ JUDGE    read each source, score vs the shared insight set (lexical or semantic)
   ├ CONDUCT  verify-yield → Elo Tournament-of-Ideas → opportunity-cost budget dispatch
   └ SCORE    insight-recall of what was selected  vs  what was achievable
              → budget_cost = the recall your budget cost you, per insight

Corpus gap vs budget gap — separated

Every insight reports achievable and obtained, distinguishing two failures that look identical elsewhere:

achievable:0 — no source in your corpus could ever satisfy it. Fetch more.
achievable:1, obtained:0 — a source had it, your budget dropped it. Spend more.

Honest state

This is L2. Two caveats stand:

1. A count model, not a neural cross-encoder. The semantic judge is real and measured, but it is a PPMI co-occurrence model built over QA-benchmark prose — specialist research vocabulary is out-of-vocabulary and scores 0. That OOV ceiling is the most likely reason the measured gain is +20% rather than larger.

2. Model-bound synthesis. Per arXiv:2604.18805, base-model strength explains 41.4% of performance variance vs 1.5% for framework design, so pure-sovereign synthesis is model-bound. Where Nishi genuinely exceeds: sovereignty, determinism/reproducibility, zero-egress, verification rigor, and opportunity-cost-aware compute.

Remaining

A true agent-level benchmark run needs source text at scale — publisher sites bot-block, so the paths are the HuggingFace dataset mirror or the Docker enterprise half. Separately, sovereign LLM synthesis stays gated on the no-float KV-cache speed rung.

Sovereign build: authored in NishiLang, compiled nx_cc→nxasm (no gcc), shipped over the Nishi mgmt API, served on the sovereign edge. Owner: librarian · lane: deep-research-sota · 2026-07-20.