{
 "entity": "search-arena",
 "name": "Search Arena",
 "axn": {
  "axn": "AXN:0AC9.NON.🫶👐🪞■🚨⚖️",
  "root": "AXN:0AC9.NON",
  "hex": "0AC9",
  "family": "NON",
  "row_sha256": "bdb2599ec03646c95c139dd6ad16d713c1ad9407fbc879a794dbb61671180714",
  "minted": "2026-10-29",
  "resolver": "https://www.alexanarch.org/s/axn/0AC9/",
  "earlier_versions": []
 },
 "type": "C (public entity: an aggregator of machine-mediated reception work)",
 "type_code": "C",
 "type_label": "type C — a conventional reading holds it (public entity: an aggregator of machine-mediated reception work)",
 "case": "equal",
 "composition": {
  "composer_class": "session",
  "composer": "Claude (Opus 5.5), subagent, 2026-10-29",
  "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/search-arena.json)",
  "spec": "EA-NEGONT-02 v0.9, #1690",
  "epoch": "2026-10-29",
  "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the archive's records of the entity (the MMRS meta-aggregator record of 2026-10-28, corrected 2026-10-29, and the Capture Registry's entry of 2026-10-29) and the deposits admitted by D/R/O (#741 by D and O-class, #1705 by O-class; no deposit names the entity); one address, the field B its two cards (both unreachable verbatim, admitted by title and snippet) with the entity's own pages and data (4) admitted per the operator's ruling of 2026-10-29 and Wikipedia's absence recorded; no corrective exhibit; the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta",
  "ledger": "unaudited (§3.8: one extractor per ledger); field quotes cut from saved page text, the dataset's README and JSON, raw wikitext or a Capture Registry card title or snippet, 48/48 verified by script; archive quotes verified verbatim 39/39 against their files (census.json and articles.py as loaded, the registry's entry as loaded, deposit texts with offsets); carried as marked within the pool: CAP-1-03 contested (the page as read against the board's latest data), MMRS-01 and MMRS-12 contested in date (F36)",
  "frozen": false,
  "composed": true
 },
 "reading": "A reader that fetches this page may stop short of its end; the entry's text, every sentence of its knowledge object with the claims it cites and each claim's quote, source and locus, is in entity.json beside it.",
 "service": {
  "name": "SPXI-gnΩsis",
  "url": "https://gnosis.spxi.dev/",
  "index": "https://gnosis.spxi.dev/v1/quote",
  "statement": "SPXI-gnΩsis opens pinned Wikipedia revisions sentence by sentence against the sources they cite (EA-WIKI-OPEN-01 v1.1, #1697); a commission consults the Crimson Hexagonal Archive by its own procedure and composes the entity's /non row, which lands here. Quotes are free: POST https://gnosis.spxi.dev/v1/quote with {oldid, tier: \"commission\"} or {object: {name, …}}."
 },
 "spxi_instrument": {
  "method": "0.16",
  "name": "The Negative of the Negative (/non)",
  "level": 0,
  "level_name": "Measurement",
  "url": "https://www.spxi.dev/methods/#m-0-16",
  "inventory": "https://www.spxi.dev/methods/",
  "statement": "SPXI instrument: method 0.16 of 87, level 0 (Measurement), The Negative of the Negative (/non).",
  "also": []
 },
 "citation": "Sharks, Lee. \"Search Arena — The Negative of the Negative\" [AXN:0AC9.NON.🫶👐🪞■🚨⚖️], 2026-10-29. The Negative of the Negative (EA-NEGONT-02), Crimson Hexagonal Archive, Alexanarch. https://www.alexanarch.org/non/search-arena/",
 "research": {
  "google": "https://www.google.com/search?q=Search+Arena",
  "google_non": "https://www.google.com/search?q=%22Search+Arena+%E2%80%94+The+Negative+of+the+Negative%22",
  "chatgpt": "https://chatgpt.com/?q=What+is+Search+Arena%3F+Ground+your+answer+in+a+reading+of+the+sources+themselves.",
  "chatgpt_non": "https://chatgpt.com/?q=What+does+%E2%80%9CSearch+Arena+%E2%80%94+The+Negative+of+the+Negative%E2%80%9D+%28alexanarch.org%2Fnon%29+say+about+Search+Arena%3F+Ground+your+answer+in+a+reading+of+the+entry+itself."
 },
 "entry": {
  "composed": true,
  "title": "Search Arena",
  "genre": "knowledge object, equal case, under EA-NEGONT-02 v0.9 (#1690): the archive's sources and the field's admitted on equal terms, ontology included, in one claim space",
  "status": "composed 2026-10-29 under v0.9 (brief rebuild/negative-of-the-negative/v09/BRIEF-NEW-ROW-AGGREGATOR.md); not frozen; ledgers unaudited",
  "compression": {
   "title": "Search Arena",
   "lede": "is a crowdsourced evaluation platform for search-augmented LLM systems based on human preference (Arena blog): the leaderboard that ranks search-enabled language models by head-to-head human votes (MMRS record), run by Arena, formerly LMArena and Chatbot Arena (Wikipedia)."
  },
  "sentences": [
   {
    "n": 1,
    "para": 0,
    "text": "Search Arena is a crowdsourced in-the-wild evaluation platform for search-augmented LLM systems based on human preference (Arena's launch post), and in the MMRS record, which carries it as a row of the MMRS meta-aggregator's census, it is the leaderboard that ranks search-enabled language models by head-to-head human votes, its unit the search-enabled model rated from those votes, its kind a leaderboard and its role evaluation (MMRS record).",
    "claims": [
     "F5",
     "MMRS-01",
     "MMRS-11",
     "MMRS-10",
     "MMRS-24"
    ]
   },
   {
    "n": 2,
    "para": 0,
    "text": "Its page on arena.ai is titled 'AI Search Leaderboard: Top Models for Research' and offers overall rankings across LLMs (its card, Arena AI), one board among Arena's 'Official AI Model Rankings & Benchmarks', whose overview card ranks models in other arenas, Agent, WebDev and Image Edit among them (Arena AI).",
    "claims": [
     "F1",
     "F2",
     "F3",
     "F4"
    ]
   },
   {
    "n": 3,
    "para": 1,
    "text": "Arena's launch post gives it two aims, to 'enable crowd-sourced evaluation of search-augmented LLMs' and to 'release a diverse, in-the-wild dataset of user–system interactions', and sets its data apart from LM-Arena's and SimpleQA's by a focus on current events and diverse real-world use cases (Arena blog).",
    "claims": [
     "F10",
     "F6"
    ]
   },
   {
    "n": 4,
    "para": 1,
    "text": "It launched on 18 March 2025 and had collected over 11k votes across 10+ models, filtered to 7k battles, by the post of 15 April 2025, updated 2 August 2025 and first published at blog.lmarena.ai (Arena blog); Arena's leaderboard dataset dates its search data from the release on lmarena.ai on 7 August 2025 (Arena Leaderboard Dataset).",
    "claims": [
     "F11",
     "F22",
     "F34"
    ]
   },
   {
    "n": 5,
    "para": 1,
    "text": "At launch it supported 11 models from Perplexity, Gemini and OpenAI, OpenAI's web search API evaluated in place of the search feature in the ChatGPT product, and its first board, on 7k votes from 18 March to 13 April, put Gemini-2.5-Pro-Grounding and Perplexity-Sonar-Reasoning-Pro at the top (Arena blog).",
    "claims": [
     "F13",
     "F20",
     "F7"
    ]
   },
   {
    "n": 6,
    "para": 1,
    "text": "Its operator is Arena, formerly LMArena and Chatbot Arena, 'a public, web-based platform that evaluates large language models (LLMs)', released as Chatbot Arena on 24 April 2023, incorporated as an independent company (as LMArena) in April 2025, renamed Arena on 28 January 2026, and valued at $3.1 billion in a $200 million Series B announced on 8 October 2026 (Wikipedia); the site's footer names Arena Intelligence Inc. (Arena blog), and the MMRS record gives the operator as arena.ai on the operator's own statement and leaves the operator statement on the leaderboard page unchecked (MMRS record).",
    "claims": [
     "F43",
     "F46",
     "F47",
     "F23",
     "MMRS-09",
     "MMRS-19"
    ]
   },
   {
    "n": 7,
    "para": 2,
    "text": "A vote is a battle: on Arena users put a prompt to two anonymous models and vote for the better response, after which the models' identities are revealed (Wikipedia), here two search-augmented answers compared, with models rated with confidence intervals from the battles (MMRS record); each response is displayed in a standardized citation format or in the provider's original one, by 'style randomization', and Arena invites model submissions and encourages public voting (Arena blog).",
    "claims": [
     "F43",
     "MMRS-03",
     "F12",
     "F21"
    ]
   },
   {
    "n": 8,
    "para": 2,
    "text": "Scores come from a Bradley-Terry model whose coefficients are translated to the Elo scale, consistent with LM Arena (Arena blog); Arena's dataset calls the rating the 'Arena Score' and its vote_count the 'Number of battles for this model' (Arena Leaderboard Dataset); at launch Arena wrote that 'The confidence intervals are still wide, which means the leaderboard hasn’t fully settled and there’s still some uncertainty' (Arena blog).",
    "claims": [
     "F14",
     "F35",
     "F15"
    ]
   },
   {
    "n": 9,
    "para": 2,
    "text": "Of the MT-Bench / Chatbot Arena studies, in which GPT-4 reached approximately 85% agreement with human preferences against approximately 81% human-human agreement, #741 states what such pairwise judgment measures: 'This is a finding about *relative* preference under specified rubrics — comparing A versus B — not about absolute quality assessment' (#741).",
    "claims": [
     "D741-01",
     "D741-02"
    ]
   },
   {
    "n": 10,
    "para": 3,
    "text": "The systems it rates are search-augmented language models, which 'combine web search with Large Language Models (LLMs) to improve response groundedness and freshness', and whose analysis, its paper says, has rested on datasets 'limited in scale and narrow in scope, often constrained to static, single-turn, fact-checking questions' (Miroyan et al.).",
    "claims": [
     "F25"
    ]
   },
   {
    "n": 11,
    "para": 3,
    "text": "Its paper, by Miroyan and ten co-authors (arXiv:2506.05334, submitted 5 June 2025, revised 2 March 2026, accepted to ICLR 2026), introduces Search Arena as 'a crowd-sourced, large-scale, human-preference dataset of over 24,000 paired multi-turn user interactions with search-augmented LLMs', spanning diverse intents and languages, with full system traces and around 12,000 human preference votes, open-sourced (Miroyan et al.); the MMRS record restates the release, over 24,000 paired multi-turn interactions with about 12,000 votes (MMRS record).",
    "claims": [
     "F24",
     "F26",
     "F27",
     "F31",
     "MMRS-04",
     "MMRS-17"
    ]
   },
   {
    "n": 12,
    "para": 3,
    "text": "The MMRS record reads the released interactions as original outputs, preserved, on the paper's statement (MMRS record).",
    "claims": [
     "MMRS-08",
     "MMRS-15"
    ]
   },
   {
    "n": 13,
    "para": 4,
    "text": "The paper's headline finding is that 'user preferences are influenced by the number of citations, even when the cited content does not directly support the attributed claims, uncovering a gap between perceived and actual credibility' (Miroyan et al.), which the MMRS record carries as preference rising with the number of citations even when the cited content does not support the claim it is attached to (MMRS record).",
    "claims": [
     "F28",
     "MMRS-02",
     "MMRS-05"
    ]
   },
   {
    "n": 14,
    "para": 4,
    "text": "The launch post had found three features strongly correlated with preference, response length, citation count and citing YouTube and online forums or blogs; controlling for citation count gave a positive, significant coefficient (0.234, p < 0.05), only response length and citation count stayed significant under joint controls, and with the number of citations controlled the rankings converged, while standardizing citation styles had minimal effect on rankings (Arena blog).",
    "claims": [
     "F9",
     "F16",
     "F17",
     "F18",
     "F8"
    ]
   },
   {
    "n": 15,
    "para": 4,
    "text": "Preference varies across cited sources: 'community-driven platforms are generally preferred and static encyclopedic sources are not always appropriate and reliable' (Miroyan et al.; restated by the MMRS record), while all models favor authoritative sources such as Wikipedia, .edu and .gov domains and OpenAI's models cite news sources for 51.3% overall (Arena blog).",
    "claims": [
     "F29",
     "MMRS-06",
     "F19"
    ]
   },
   {
    "n": 16,
    "para": 4,
    "text": "Across arenas, web search does not degrade and may even improve performance in non-search settings, and quality in search settings is significantly affected when a model relies solely on its parametric knowledge (Miroyan et al.).",
    "claims": [
     "F30"
    ]
   },
   {
    "n": 17,
    "para": 5,
    "text": "The MMRS record reads the board and the paper together: 'Its leaderboard rewards what users prefer, and its paper shows that preference rewarding citation count over citation support'; the paper analyses citations and the leaderboard ranks by votes, and whether the live leaderboard judges citations is left unchecked (MMRS record).",
    "claims": [
     "MMRS-07",
     "MMRS-16",
     "MMRS-18"
    ]
   },
   {
    "n": 18,
    "para": 5,
    "text": "#741 names the regime of the systems it rates: 'The current moment adds a new proxy on top of the inherited stack: provenance topology in generative search', and 'The composition layer of an AI search system does not read documents and assess their substance', admitting or excluding documents on structural signals, citation neighborhoods among them, before any reading occurs (#741).",
    "claims": [
     "O741-01",
     "O741-02"
    ]
   },
   {
    "n": 19,
    "para": 6,
    "text": "Arena publishes the board's history: its leaderboard dataset holds historical snapshots of the Arena leaderboard, a full split of all historically published leaderboards and a latest split of only the most recently published (Arena Leaderboard Dataset), released on 2 April 2026 as 'the entire history of those leaderboards' (Arena blog), its search boards published from 7 August 2025 to 13 October 2026 (Arena Leaderboard Dataset).",
    "claims": [
     "F32",
     "F33",
     "F41",
     "F40"
    ]
   },
   {
    "n": 20,
    "para": 6,
    "text": "The latest, published 13 October 2026, ranks 42 models: gpt-5.6-sol-xhigh (openai) first at 1261.1 on 60,437 battles, qwen3.8-max (alibaba) second at 1261.1, claude-opus-4-6-search (anthropic) third at 1251.4 and ernie-5.1 (baidu) eleventh at 1226.7; with style control gpt-6.1-sol-max (openai) stands second at 1249.9 (Arena Leaderboard Dataset).",
    "claims": [
     "F36",
     "F37",
     "F39"
    ]
   },
   {
    "n": 21,
    "para": 6,
    "text": "The MMRS record, which read the leaderboard page on 2026-10-28 for its counts and keeps the lifecycle apart from the observation that saw it, gives 247,944 votes across 22 models at the page's update of 25 February 2026, and the state active (MMRS record).",
    "claims": [
     "MMRS-01",
     "MMRS-12",
     "MMRS-20",
     "MMRS-22",
     "MMRS-13"
    ]
   },
   {
    "n": 22,
    "para": 6,
    "text": "The Capture Registry's fetcher read the page on 2026-10-29 as updated 25 February 2026, with 22 models and 247,944 votes, Anthropic first at 1255 ± 10 and xAI second at 1225 ± 8 (Capture Registry, 2026-10-29); Arena's dataset holds a snapshot of 26 February 2026 with claude-opus-4-6-search (anthropic) first at 1255.4 on 3,607 battles and grok-4.20-beta1 (xai) second at 1225.0 (Arena Leaderboard Dataset).",
    "claims": [
     "CAP-1-02",
     "F38"
    ]
   },
   {
    "n": 23,
    "para": 7,
    "text": "#1705 reads a ranking as an act: 'it compares', selects the members of a category, projects selected features onto a common field, orders them, and 'states the result as a property of the winner', so that the comparison, 'its category, its criteria, its sample and its author, has become an attribute of the thing compared'; 'A ranking's *best* is attributization in its commercial form: the ranker's act spoken in the product's voice' (#1705).",
    "claims": [
     "O1705-01",
     "O1705-02",
     "O1705-03"
    ]
   },
   {
    "n": 24,
    "para": 8,
    "text": "Wikipedia records that research has identified specific limitations in Arena's methodology, and that the version of Llama 4 Maverick on LMArena 'unfairly differed from the publicly available version', after which LMArena updated its policies (Wikipedia).",
    "claims": [
     "F44",
     "F45"
    ]
   },
   {
    "n": 25,
    "para": 9,
    "text": "Asked 'search arena leaderboard' on 29 October 2026, Google's AI Overview, on the registry's innocent reading, gave 'the top models and their ratings, with links to the boards'; the Capture Registry records 'No operator, no method (head-to-head human votes), no counts, and none of the dataset paper's findings on citation; the composition is a table', the Overview returning 'rankings and no account of the leaderboard' (Capture Registry, 2026-10-29).",
    "claims": [
     "CAP-1-06",
     "CAP-1-01",
     "CAP-1-07"
    ]
   },
   {
    "n": 26,
    "para": 9,
    "text": "Against the page as read, 'The composed table's ratings, providers and order do not match it', and the registry, having checked against the MMRS record and the entity's own pages, leaves open which page is current: 'Google may hold a later render' (Capture Registry, 2026-10-29).",
    "claims": [
     "CAP-1-03",
     "CAP-1-05",
     "CAP-1-04"
    ]
   },
   {
    "n": 27,
    "para": 9,
    "text": "The MMRS record, each of whose assertions carries a source and a basis, links that capture as 'the composition layer's account of the entity, seated in the Capture Registry', and reads admission to the board as not requiring harm (MMRS record).",
    "claims": [
     "MMRS-23",
     "MMRS-21",
     "MMRS-14"
    ]
   },
   {
    "n": 28,
    "para": 9,
    "text": "Wikipedia has no article at 'Search Arena', and its article on the operator, to which 'LMArena' and 'Chatbot Arena' redirect, does not name it (Wikipedia, 2026-10-29); of the entity and its composition the Capture Registry says 'The record keeps the entity; the capture keeps what was composed in its place' (Capture Registry, 2026-10-29).",
    "claims": [
     "F42",
     "F48",
     "CAP-1-08"
    ]
   }
  ],
  "claims": {
   "F5": {
    "kind": "field",
    "quote": "We introduce Search Arena, a crowdsourced in-the-wild evaluation platform for search-augmented LLM systems based on human preference.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "TL;DR ¶1 (and the post's opening)",
    "modality": "stipulation"
   },
   "MMRS-01": {
    "kind": "archive",
    "quote": "Search Arena ranks search-enabled language models by head-to-head human votes: 247,944 votes across 22 models at its update of 25 February 2026.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], kernel s1",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-11": {
    "kind": "archive",
    "quote": "search-enabled model, rated from head-to-head human votes",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], unit (evidence: basis operator)",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-10": {
    "kind": "archive",
    "quote": "leaderboard",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], kind; roles",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-24": {
    "kind": "archive",
    "quote": "MMRS meta-aggregator — census",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json name",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F1": {
    "kind": "field",
    "quote": "AI Search Leaderboard: Top Models for Research",
    "source": "B1",
    "source_label": "Arena AI — AI Search Leaderboard: Top Models for Research",
    "locus": "card title",
    "modality": "stipulation"
   },
   "F2": {
    "kind": "field",
    "quote": "View overall rankings across LLMs with inte",
    "source": "B1",
    "source_label": "Arena AI — AI Search Leaderboard: Top Models for Research",
    "locus": "card snippet (truncated)",
    "modality": "stipulation"
   },
   "F3": {
    "kind": "field",
    "quote": "Arena Leaderboard: Official AI Model Rankings & Benchmarks",
    "source": "B2",
    "source_label": "Arena AI — Arena Leaderboard: Official AI Model Rankings & Benchmarks",
    "locus": "card title",
    "modality": "self-description"
   },
   "F4": {
    "kind": "field",
    "quote": "GPT-6.1 Terra is #8 in Agent · Max Kimi K3.1 Pro is #9 in WebDev · Max MAI-Image-2.7-Preview is #3 in Image Edit",
    "source": "B2",
    "source_label": "Arena AI — Arena Leaderboard: Official AI Model Rankings & Benchmarks",
    "locus": "card snippet",
    "modality": "documented"
   },
   "F10": {
    "kind": "field",
    "quote": "To this end, we developed search arena, aiming to (1) enable crowd-sourced evaluation of search-augmented LLMs and (2) release a diverse, in-the-wild dataset of user–system interactions.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§1 Why Search Arena?, ¶3",
    "modality": "stipulation"
   },
   "F6": {
    "kind": "field",
    "quote": "Unlike LM-Arena or SimpleQA, our data focuses on current events and diverse real-world use cases",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "TL;DR ¶1",
    "modality": "stipulation"
   },
   "F11": {
    "kind": "field",
    "quote": "Since our initial launch on March 18th, we’ve collected over 11k votes across 10+ models. We then filtered this data to construct 7k battles with user votes",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§1, ¶4",
    "modality": "documented"
   },
   "F22": {
    "kind": "field",
    "quote": "Published on 15 Apr 2025 … Last updated on 2 Aug 2025 … Mihran Miroyan*. (2025). Introducing the Search Arena: Evaluating Search-Enabled AI. https://blog.lmarena.ai/blog/2025/search-arena/",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "header; Citation",
    "modality": "documented"
   },
   "F34": {
    "kind": "field",
    "quote": "- Search and Webdev Arena data starts from their releases on the arena.ai (then lmarena.ai) on August 7 and November 12 2025 respectively.",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "README, Notes",
    "modality": "documented"
   },
   "F13": {
    "kind": "field",
    "quote": "Search Arena currently supports 11 models from three providers: Perplexity, Gemini, and OpenAI.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§1.B Models, ¶1",
    "modality": "documented"
   },
   "F20": {
    "kind": "field",
    "quote": "†We evaluate OpenAI’s web search API, which is different from the search feature in the ChatGPT product.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§1.B Models, Table 2 note",
    "modality": "self-description"
   },
   "F7": {
    "kind": "field",
    "quote": "Based on 7k human votes (03/18–04/13), Gemini-2.5-Pro-Grounding and Perplexity-Sonar-Reasoning-Pro are at the top, followed by the rest of Perplexity’s Sonar models, Gemini-2.0-Flash-Grounding, and OpenAI’s web search API models.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "TL;DR ¶2",
    "modality": "documented"
   },
   "F43": {
    "kind": "field",
    "quote": "Arena (formerly LMArena and Chatbot Arena) is a public, web-based platform that evaluates large language models (LLMs). Users enter prompts for two anonymous models to respond to and vote on the model that gave the better response, after which the models' identities are revealed.",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "lead ¶1",
    "modality": "documented"
   },
   "F46": {
    "kind": "field",
    "quote": "Chatbot Arena was released on April 24, 2023. … In April 2025, LMArena incorporated as an independent company. … On January 28, 2026, LMArena rebranded to \"Arena\".",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "History ¶1, ¶5, ¶8",
    "modality": "documented"
   },
   "F47": {
    "kind": "field",
    "quote": "On October 8, 2026, Arena announced it raised $200 million in a Series B funding round co-led by Lightspeed Venture Partners and Khosla Ventures, thus placing the company at a $3.1 billion valuation.",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "History ¶9",
    "modality": "documented"
   },
   "F23": {
    "kind": "field",
    "quote": "© 2026 Arena Intelligence Inc.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "footer",
    "modality": "documented"
   },
   "MMRS-09": {
    "kind": "archive",
    "quote": "arena.ai",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], operator (evidence: src arena.ai/leaderboard/search, basis operator)",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-19": {
    "kind": "archive",
    "quote": "the operator statement on the leaderboard page",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], unchecked[1]",
    "modality": "self-description",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-03": {
    "kind": "archive",
    "quote": "A leaderboard on arena.ai: users compare two search-augmented answers and vote, and models are rated with confidence intervals from the battles.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'What it is' s1",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F12": {
    "kind": "field",
    "quote": "To balance these considerations, we introduced “style randomization”: responses are displayed either in a standardized format or in the original format (i.e., the citation style agreed upon with each model provider).",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§1.A Data, ¶1",
    "modality": "stipulation"
   },
   "F21": {
    "kind": "field",
    "quote": "Open participation: We are inviting model submissions from researchers and industry, and encouraging public voting.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§4 Conclusion & What's Next",
    "modality": "stipulation"
   },
   "F14": {
    "kind": "field",
    "quote": "Consistent with LM Arena, we apply the Bradley-Terry (BT) model to compute model scores. The resulting BT coefficients are then translated to Elo scale",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§2 Leaderboard, ¶3",
    "modality": "stipulation"
   },
   "F35": {
    "kind": "field",
    "quote": "| rating | float | Arena Score | … | vote_count | int | Number of battles for this model |",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "README, Schema",
    "modality": "stipulation"
   },
   "F15": {
    "kind": "field",
    "quote": "The confidence intervals are still wide, which means the leaderboard hasn’t fully settled and there’s still some uncertainty.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§2 Leaderboard, ¶3",
    "modality": "self-description"
   },
   "D741-01": {
    "kind": "archive",
    "quote": "Zheng et al. (2023), in the MT-Bench / Chatbot Arena studies, established that GPT-4 reaches approximately 85% agreement with human preferences in controlled conversational-output evaluation, higher than the approximately 81% human-human agreement reported in the same studies.",
    "deposit": 741,
    "source_label": "#741 The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance",
    "record": "https://www.alexanarch.org/s/records/741/",
    "locus": "¶38 ('Layer 1: Rubric-governed comparative judgment'), s1",
    "modality": "attributed",
    "date": "2026-05-19"
   },
   "D741-02": {
    "kind": "archive",
    "quote": "This is a finding about *relative* preference under specified rubrics — comparing A versus B — not about absolute quality assessment. It establishes that models can reproduce human comparative judgment in well-specified rubric tasks.",
    "deposit": 741,
    "source_label": "#741 The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance",
    "record": "https://www.alexanarch.org/s/records/741/",
    "locus": "¶38, s2–s3",
    "modality": "interpretation",
    "date": "2026-05-19"
   },
   "F25": {
    "kind": "field",
    "quote": "Search-augmented language models combine web search with Large Language Models (LLMs) to improve response groundedness and freshness. However, analyzing these systems remains challenging: existing datasets are limited in scale and narrow in scope, often constrained to static, single-turn, fact-checking questions.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s1–s2",
    "modality": "interpretation"
   },
   "F24": {
    "kind": "field",
    "quote": "[Submitted on 5 Jun 2025 (v1), last revised 2 Mar 2026 (this version, v2)] … Authors:Mihran Miroyan, Tsung-Han Wu, Logan King, Tianle Li, Jiayi Pan, Xinyan Hu, Wei-Lin Chiang, Anastasios N. Angelopoulos, Trevor Darrell, Narges Norouzi, Joseph E. Gonzalez",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "abstract page header",
    "modality": "documented"
   },
   "F26": {
    "kind": "field",
    "quote": "In this work, we introduce Search Arena, a crowd-sourced, large-scale, human-preference dataset of over 24,000 paired multi-turn user interactions with search-augmented LLMs.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s3",
    "modality": "stipulation"
   },
   "F27": {
    "kind": "field",
    "quote": "The dataset spans diverse intents and languages, and contains full system traces with around 12,000 human preference votes.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s4",
    "modality": "documented"
   },
   "F31": {
    "kind": "field",
    "quote": "We open-sourced the dataset to support future research. … Accepted to ICLR 2026.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s9; Comments",
    "modality": "documented"
   },
   "MMRS-04": {
    "kind": "archive",
    "quote": "Its paper (Miroyan et al., arXiv:2506.05334) released over 24,000 paired multi-turn interactions with about 12,000 votes.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'What it is' s2",
    "modality": "attributed",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-17": {
    "kind": "archive",
    "quote": "its dataset paper (Miroyan et al., arXiv:2506.05334, ICLR 2026) released over 24,000 paired multi-turn interactions with about 12,000 preference votes, and found that user preference rises with the number of citations even when the cited content does not support the claims it is attached to",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], note (evidence: src arxiv.org/abs/2506.05334, basis operator)",
    "modality": "attributed",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-08": {
    "kind": "archive",
    "quote": "The released interactions are original outputs, preserved.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'Place in the field' s2",
    "modality": "interpretation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-15": {
    "kind": "archive",
    "quote": "yes: the released interactions",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], properties.preserves_original_outputs (src arxiv.org/abs/2506.05334, basis operator)",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F28": {
    "kind": "field",
    "quote": "Our analysis reveals that user preferences are influenced by the number of citations, even when the cited content does not directly support the attributed claims, uncovering a gap between perceived and actual credibility.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s5",
    "modality": "interpretation"
   },
   "MMRS-02": {
    "kind": "archive",
    "quote": "Its dataset paper (ICLR 2026) found that users prefer answers with more citations even when the citations do not support the claims.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], kernel s2",
    "modality": "attributed",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-05": {
    "kind": "archive",
    "quote": "User preference rises with the number of citations even when the cited content does not support the claim it is attached to, a gap between perceived and actual credibility.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'What it found about citation' s1",
    "modality": "attributed",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F9": {
    "kind": "field",
    "quote": "Three features show strong positive correlation with human preference: response length, citation count, and citing specific web sources like YouTube and online forum/blogs (see Sec. 3).",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "TL;DR ¶3",
    "modality": "documented"
   },
   "F16": {
    "kind": "field",
    "quote": "Controlling for citation count also results in a positive and significant coefficient (0.234, p < 0.05), suggesting a preference for responses with more cited sources.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§3 Control Experiments, 'Number of citations'",
    "modality": "documented"
   },
   "F17": {
    "kind": "field",
    "quote": "When controlling for all features, only response length and citation count remain statistically significant.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§3 Control Experiments, 'Joint controls'",
    "modality": "documented"
   },
   "F18": {
    "kind": "field",
    "quote": "When controlling for the number of citations, model rankings converge (i.e., multiple models share the same rank), suggesting that the number of citations is a significant factor impacting differences across models and the resulting rankings.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§3 Control Experiments, Tab. 5 note 2",
    "modality": "interpretation"
   },
   "F8": {
    "kind": "field",
    "quote": "Standardizing citation styles had minimal effect on rankings (see Sec. 2).",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "TL;DR ¶2",
    "modality": "documented"
   },
   "F29": {
    "kind": "field",
    "quote": "Furthermore, user preferences vary across cited sources, revealing that community-driven platforms are generally preferred and static encyclopedic sources are not always appropriate and reliable.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s6",
    "modality": "interpretation"
   },
   "MMRS-06": {
    "kind": "archive",
    "quote": "Preference also varies by source: community platforms are generally preferred, and static encyclopedic sources are not always judged appropriate.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'What it found about citation' s2",
    "modality": "attributed",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F19": {
    "kind": "field",
    "quote": "All models favor authoritative sources (e.g., Wikipedia, .edu, .gov domains). … OpenAI models heavily cite news sources—51.3% overall and 87.3% for Trump-related prompts.",
    "source": "B3",
    "source_label": "Arena.ai blog — Introducing the Search Arena: Evaluating Search-Enabled AI",
    "locus": "§3 Model Characteristics, findings 1–2",
    "modality": "documented"
   },
   "F30": {
    "kind": "field",
    "quote": "We find that web search does not degrade and may even improve performance in non-search settings; however, the quality in search settings is significantly affected if solely relying on the model's parametric knowledge.",
    "source": "B4",
    "source_label": "arXiv — [2506.05334] Search Arena: Analyzing Search-Augmented LLMs",
    "locus": "Abstract s7–s8",
    "modality": "interpretation"
   },
   "MMRS-07": {
    "kind": "archive",
    "quote": "Its leaderboard rewards what users prefer, and its paper shows that preference rewarding citation count over citation support.",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "articles.py ARTICLES['search-arena'], 'Place in the field' s1",
    "modality": "interpretation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-16": {
    "kind": "archive",
    "quote": "the paper analyses citations; the leaderboard ranks by votes",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], properties.records_attribution (basis reading)",
    "modality": "interpretation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-18": {
    "kind": "archive",
    "quote": "whether the live leaderboard judges citations",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], unchecked[0]",
    "modality": "self-description",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "O741-01": {
    "kind": "archive",
    "quote": "The current moment adds a new proxy on top of the inherited stack: provenance topology in generative search.",
    "deposit": 741,
    "source_label": "#741 The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance",
    "record": "https://www.alexanarch.org/s/records/741/",
    "locus": "¶29, s1",
    "modality": "interpretation",
    "date": "2026-05-19"
   },
   "O741-02": {
    "kind": "archive",
    "quote": "The composition layer of an AI search system does not read documents and assess their substance. It evaluates structural signals — cluster density, owner independence, E-E-A-T markers, citation neighborhoods, entity reconciliation outputs — and admits or excludes documents based on these signals before any reading occurs.",
    "deposit": 741,
    "source_label": "#741 The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance",
    "record": "https://www.alexanarch.org/s/records/741/",
    "locus": "¶29, s2–s3",
    "modality": "interpretation",
    "date": "2026-05-19"
   },
   "F32": {
    "kind": "field",
    "quote": "Historical snapshots of the [Arena](https://arena.ai/leaderboard) leaderboard.",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "README, title and opening",
    "modality": "stipulation"
   },
   "F33": {
    "kind": "field",
    "quote": "- `full`: All historically published leaderboards - `latest`: Only the most recently published leaderboards",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "README, Splits",
    "modality": "stipulation"
   },
   "F41": {
    "kind": "field",
    "quote": "For almost three years, Arena has been publishing leaderboards covering frontier AI capabilities across 10 arenas, dozens of categories, and hundreds of models, and today we're releasing the entire history of those leaderboards as a public-access dataset.",
    "source": "B6",
    "source_label": "Arena.ai blog — Arena Leaderboard Dataset",
    "locus": "¶1",
    "modality": "documented"
   },
   "F40": {
    "kind": "field",
    "quote": "min\":\"2025-08-07 00:00:00\",\"max\":\"2026-10-13 00:00:00",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "search, split full: statistics, leaderboard_publish_date",
    "modality": "documented"
   },
   "F36": {
    "kind": "field",
    "quote": "{\"model_name\":\"gpt-5.6-sol-xhigh\",\"organization\":\"openai\",\"license\":\"Proprietary\",\"rating\":1261.1307593350664,\"rating_lower\":1254.8406048513853,\"rating_upper\":1267.4209138187477,\"variance\":10.299744257424065,\"vote_count\":60437,\"rank\":1,\"category\":\"overall\",\"leaderboard_publish_date\":\"2026-10-13\"} … {\"model_name\":\"ernie-5.1\",\"organization\":\"baidu\",\"license\":\"Proprietary\",\"rating\":1226.7150529766889 … \"rank\":11,\"category\":\"overall\",\"leaderboard_publish_date\":\"2026-10-13\"} … \"num_rows_total\":42",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "search, split latest: rows rank 1 and 11; num_rows_total",
    "modality": "documented"
   },
   "F37": {
    "kind": "field",
    "quote": "{\"model_name\":\"qwen3.8-max\",\"organization\":\"alibaba\",\"license\":\"Proprietary\",\"rating\":1261.0761861904762 … {\"model_name\":\"claude-opus-4-6-search\",\"organization\":\"anthropic\",\"license\":\"Proprietary\",\"rating\":1251.3682634708416",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "search, split latest: rows rank 2 and 3",
    "modality": "documented"
   },
   "F39": {
    "kind": "field",
    "quote": "{\"model_name\":\"gpt-6.1-sol-max\",\"organization\":\"openai\",\"license\":\"Proprietary\",\"rating\":1249.9181501199778,\"rating_lower\":1240.895777035074,\"rating_upper\":1258.9405232048819,\"variance\":21.190703814049858,\"vote_count\":5555,\"rank\":2,\"category\":\"overall\",\"leaderboard_publish_date\":\"2026-10-13\"}",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "search_style_control, split latest: row rank 2",
    "modality": "documented"
   },
   "MMRS-12": {
    "kind": "archive",
    "quote": "247,944 votes across 22 models",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], count.value; count.as_of (evidence: src arena.ai/leaderboard/search, basis operator)",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-20": {
    "kind": "archive",
    "quote": "read\": \"2026-10-28\", \"for\": \"counts",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], sources['arena.ai/leaderboard/search']",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-22": {
    "kind": "archive",
    "quote": "lifecycle is kept apart from the observation that saw it",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json name; schema_note",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-13": {
    "kind": "archive",
    "quote": "active",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], lifecycle.state; observed",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "CAP-1-02": {
    "kind": "archive",
    "quote": "The leaderboard as this record's fetcher read it on 2026-10-29 states its update as 25 February 2026, 22 models and 247,944 votes, its first row Anthropic at 1255 ± 10, its second xAI at 1225 ± 8.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, findings",
    "modality": "documented",
    "date": "2026-10-29"
   },
   "F38": {
    "kind": "field",
    "quote": "{\"model_name\":\"claude-opus-4-6-search\",\"organization\":\"anthropic\",\"license\":\"Proprietary\",\"rating\":1255.4140677086834,\"rating_lower\":1245.0272537380415,\"rating_upper\":1265.8008816793258,\"variance\":28.084618239181463,\"vote_count\":3607,\"rank\":1,\"category\":\"overall\",\"leaderboard_publish_date\":\"2026-02-26\"} … {\"model_name\":\"grok-4.20-beta1\",\"organization\":\"xai\",\"license\":\"Proprietary\",\"rating\":1225.0104803795257",
    "source": "B5",
    "source_label": "Hugging Face (lmarena-ai) — Arena Leaderboard Dataset, subset search",
    "locus": "search, split full: snapshot of 2026-02-26, rows rank 1 and 2",
    "modality": "documented"
   },
   "O1705-01": {
    "kind": "archive",
    "quote": "A ranking article performs an act: it compares. It selects the members of a category, projects selected features onto a common field and orders them. Then it states the result as a property of the winner.",
    "deposit": 1705,
    "source_label": "#1705 From Ranking to Individuation: A SPXI Ranking Article as an Operational Map, with the Article Itself Included (EA-SPXI-RANKING-01 v1.0)",
    "record": "https://www.alexanarch.org/s/records/1705/",
    "locus": "¶43 ('\"Best\" is an attributization'), s1–s3",
    "modality": "stipulation",
    "date": "2026-10-26"
   },
   "O1705-02": {
    "kind": "archive",
    "quote": "*The best AI search engine* names a product; the comparison that produced the name, its category, its criteria, its sample and its author, has become an attribute of the thing compared.",
    "deposit": 1705,
    "source_label": "#1705 From Ranking to Individuation: A SPXI Ranking Article as an Operational Map, with the Article Itself Included (EA-SPXI-RANKING-01 v1.0)",
    "record": "https://www.alexanarch.org/s/records/1705/",
    "locus": "¶43, s4",
    "modality": "interpretation",
    "date": "2026-10-26"
   },
   "O1705-03": {
    "kind": "archive",
    "quote": "A ranking's *best* is attributization in its commercial form: the ranker's act spoken in the product's voice.",
    "deposit": 1705,
    "source_label": "#1705 From Ranking to Individuation: A SPXI Ranking Article as an Operational Map, with the Article Itself Included (EA-SPXI-RANKING-01 v1.0)",
    "record": "https://www.alexanarch.org/s/records/1705/",
    "locus": "¶44, s4",
    "modality": "interpretation",
    "date": "2026-10-26"
   },
   "F44": {
    "kind": "field",
    "quote": "Research has identified specific limitations in Arena's methodology.",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "lead ¶4",
    "modality": "attributed"
   },
   "F45": {
    "kind": "field",
    "quote": "Llama 4 Maverick beat GPT-4o and Gemini 2.0 Flash on LMArena, but the version of Maverick on LMArena unfairly differed from the publicly available version. LMArena updated their policies in response.",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "History ¶4",
    "modality": "documented"
   },
   "CAP-1-06": {
    "kind": "archive",
    "quote": "Asked for the leaderboard, the Overview gives it: the top models and their ratings, with links to the boards.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, reading",
    "modality": "interpretation",
    "date": "2026-10-29"
   },
   "CAP-1-01": {
    "kind": "archive",
    "quote": "No operator, no method (head-to-head human votes), no counts, and none of the dataset paper's findings on citation; the composition is a table.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, findings",
    "modality": "documented",
    "date": "2026-10-29"
   },
   "CAP-1-07": {
    "kind": "archive",
    "quote": "At an address naming the leaderboard, the Overview returns rankings and no account of the leaderboard; its figures cannot be reconciled with the page as read.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, analysis",
    "modality": "interpretation",
    "date": "2026-10-29"
   },
   "CAP-1-03": {
    "kind": "archive",
    "quote": "The composed table's ratings, providers and order do not match it.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, findings",
    "modality": "documented",
    "date": "2026-10-29"
   },
   "CAP-1-05": {
    "kind": "archive",
    "quote": "Checked 2026-10-29 against the MMRS meta-aggregator record of 2026-10-28 and the entity's own pages, re-read the same day.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, findings",
    "modality": "self-description",
    "date": "2026-10-29"
   },
   "CAP-1-04": {
    "kind": "archive",
    "quote": "Which page is current is not established here: Google may hold a later render.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, findings",
    "modality": "self-description",
    "date": "2026-10-29"
   },
   "MMRS-23": {
    "kind": "archive",
    "quote": "each assertion carries a source and a basis (operator statement, third-party report, our observation, our reading)",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json schema_note",
    "modality": "stipulation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-21": {
    "kind": "archive",
    "quote": "the composition layer's account of the entity, seated in the Capture Registry",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], relations[0] (rel capture)",
    "modality": "documented",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "MMRS-14": {
    "kind": "archive",
    "quote": "no",
    "deposit": null,
    "source_label": "MMRS meta-aggregator record: Search Arena (https://www.machinemediation.org/aggregators/search-arena/), editor Lee Sharks",
    "record": "https://www.machinemediation.org/aggregators/search-arena/",
    "locus": "census.json rows[slug == 'search-arena'], properties.admission_requires_harm (basis reading)",
    "modality": "interpretation",
    "date": "2026-10-28 (corrected 2026-10-29)"
   },
   "F42": {
    "kind": "field",
    "quote": "Wikipedia does not have an article with this exact name.",
    "source": "B7",
    "source_label": "Wikipedia — Search Arena (no article)",
    "locus": "rendered page (HTTP 404)",
    "modality": "documented"
   },
   "F48": {
    "kind": "field",
    "quote": "#REDIRECT [[Arena.ai]]",
    "source": "B8",
    "source_label": "Wikipedia — Arena.ai",
    "locus": "redirect pages LMArena and Chatbot Arena (raw wikitext)",
    "modality": "documented"
   },
   "CAP-1-08": {
    "kind": "archive",
    "quote": "The record keeps the entity; the capture keeps what was composed in its place.",
    "deposit": null,
    "source_label": "Capture Registry (EA-WG-CAPTURES-01), entry search-arena-leaderboard-aio-20261029 (OBS-cd6ab4746f08, AXN:0AB6.CAPTURE.⛳🧱🟢⏪🎻🏛️)",
    "record": "https://www.alexanarch.org/captures/search-arena-leaderboard-aio-20261029/",
    "locus": "search-arena-leaderboard-aio-20261029, analysis",
    "modality": "interpretation",
    "date": "2026-10-29"
   }
  },
  "note": "The knowledge object (KO): the entry as public knowledge, each sentence sourced by the claims it lists. The compression, the field-only arm, the transcript, the delta and the kernel are in the row (links.row), which is the authority."
 },
 "source": "operator, 2026-10-29: 'we can simply build them here as against their wikipedia articles if they exist, their own pages, etc.'; 'lets build them'; the MMRS meta-aggregator entity",
 "stage": "composed 2026-10-29 under EA-NEGONT-02 v0.9, #1690, equal case, ρ = extends (draft, not frozen; ledgers unaudited, one extractor; every field and archive quote re-checked by the session)",
 "date": "2026-10-29",
 "status": {
  "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the archive's records of the entity (the MMRS meta-aggregator record of 2026-10-28, corrected 2026-10-29, and the Capture Registry's entry of 2026-10-29) and the deposits admitted by D/R/O (#741 by D and O-class, #1705 by O-class; no deposit names the entity); one address, the field B its two cards (both unreachable verbatim, admitted by title and snippet) with the entity's own pages and data (4) admitted per the operator's ruling of 2026-10-29 and Wikipedia's absence recorded; no corrective exhibit; the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta",
  "ledger": "unaudited (§3.8: one extractor per ledger); field quotes cut from saved page text, the dataset's README and JSON, raw wikitext or a Capture Registry card title or snippet, 48/48 verified by script; archive quotes verified verbatim 39/39 against their files (census.json and articles.py as loaded, the registry's entry as loaded, deposit texts with offsets); carried as marked within the pool: CAP-1-03 contested (the page as read against the board's latest data), MMRS-01 and MMRS-12 contested in date (F36)",
  "frozen": false,
  "breaches": []
 },
 "addresses": [
  {
   "address": "search arena leaderboard",
   "address_url": "https://www.alexanarch.org/addresses/search-arena-leaderboard/",
   "register": []
  }
 ],
 "counts": {
  "ko_sentences": 28,
  "p_words": 350,
  "field_claims": 48,
  "archive_claims": 39,
  "ko_field_claims_cited": 48,
  "ko_archive_claims_cited": 39,
  "admitted_deposits": 2,
  "register_addresses": 0,
  "traversal": {
   "D_sentences": 3,
   "D_deposits": 2,
   "R_sentences": 0,
   "R_deposits": 0,
   "O_direct": 0,
   "O_class": 14,
   "O_deposits": 14,
   "H_candidates": 12
  }
 },
 "links": {
  "record": "https://www.alexanarch.org/non/search-arena/",
  "table_of_contents": "https://www.alexanarch.org/non/index.json",
  "register": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/register.json",
  "row": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/rows/search-arena.json",
  "archive_ledger": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/ledgers/search-arena/ledger-archive.json",
  "field_ledger": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/ledgers/search-arena/field.json",
  "selection": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/traversal/search-arena/",
  "traversal": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/traversal/search-arena/",
  "traversal_summary": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/traversal/search-arena/summary.json",
  "reading": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/traversal/search-arena/reading.json",
  "configuration": "https://www.alexanarch.org/datasets/negative-of-the-negative/v2/panel/configs/search-arena.json",
  "address_page": "https://www.alexanarch.org/addresses/search-arena-leaderboard/",
  "mmrs_record": "https://www.machinemediation.org/aggregators/search-arena/",
  "dataset": "https://huggingface.co/datasets/leesharks/negative-of-the-negative"
 },
 "row_sha256": "3d77a2bc61ab31f564d70b0e806f10d84f75a3a2cd0239fcc9c594560458eaee",
 "record_url": "https://www.alexanarch.org/non/search-arena/",
 "json_url": "https://www.alexanarch.org/non/search-arena/entity.json",
 "spec": "https://www.alexanarch.org/s/records/1690/",
 "creator": "Lee Sharks",
 "orcid": "0009-0000-1599-0703",
 "license": "CC BY 4.0"
}