{"config":"agent_tool_hallucination","id":"arena-agent-tool-hallucination","title":"Agents · tool hallucination signal","category":"Coding and agents","metric":"IPS score","direction":"higher","precision":6,"publishedAt":"2026-11-04","interval":"95% confidence interval","observationLabel":"Observations","description":"Agent Arena signal scores from the official dataset. Each signal is a separate measurement; it is not a success percentage.","limitations":["The population is the models in this source board, including variants and models without an available API price.","Ranks use the published confidence bounds: one plus the number of models whose lower bound exceeds this model’s upper bound. Shared ranks are not proof of equal performance."],"omissions":[],"rows":[{"name":"GLM-5.3-Flash","organization":"zai","score":0.008902640510079607,"low":0.00808054385911856,"high":0.009724737161040654,"observations":11871656,"sourceRank":1,"publishedAt":"2026-11-04","rank":1},{"name":"Grok 4.6 (xHigh)","organization":"xai","score":0.008902640510001447,"low":0.008080543859040416,"high":0.009724737160962478,"observations":3557471,"sourceRank":2,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6 Luna (Low)","organization":"openai","score":0.00890264050962486,"low":0.008080543858663893,"high":0.009724737160585827,"observations":2732091,"sourceRank":3,"publishedAt":"2026-11-04","rank":1},{"name":"GLM-5.3 (Max)","organization":"zai","score":0.008902640509525384,"low":0.008080543858564422,"high":0.009724737160486345,"observations":10241529,"sourceRank":4,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6.1 Sol (Max)","organization":"openai","score":0.008902640509262483,"low":0.008080543858301535,"high":0.00972473716022343,"observations":2555443,"sourceRank":5,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6 Astra (Max)","organization":"openai","score":0.008902640509077742,"low":0.008080543858116815,"high":0.009724737160038669,"observations":1194324,"sourceRank":6,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6.1 Terra (Max)","organization":"openai","score":0.008902640509020898,"low":0.008080543858059978,"high":0.009724737159981818,"observations":6051267,"sourceRank":7,"publishedAt":"2026-11-04","rank":1},{"name":"Kimi K3.1 Pro (Max)","organization":"moonshot","score":0.008902640508736681,"low":0.008080543857775815,"high":0.009724737159697548,"observations":4948298,"sourceRank":8,"publishedAt":"2026-11-04","rank":1},{"name":"DeepSeek-V4-Pro (High, 20260813)","organization":"deepseek","score":0.00890264050870826,"low":0.008080543857747372,"high":0.009724737159669147,"observations":812009,"sourceRank":9,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-5.6 Sol (xHigh)","organization":"openai","score":0.008902640508154036,"low":0.008080543857193244,"high":0.009724737159114828,"observations":6106979,"sourceRank":10,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6 Luna (Medium)","organization":"openai","score":0.00890264050813272,"low":0.008080543857171855,"high":0.009724737159093584,"observations":7544280,"sourceRank":11,"publishedAt":"2026-11-04","rank":1},{"name":"Kimi K3 (Max)","organization":"moonshot","score":0.008902326598629884,"low":0.008080229862671973,"high":0.009724423334587794,"observations":14767204,"sourceRank":12,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6 Sol (Max)","organization":"openai","score":0.008901059891449847,"low":0.00807895787511364,"high":0.009723161907786054,"observations":2813701,"sourceRank":13,"publishedAt":"2026-11-04","rank":1},{"name":"Grok 4.7 (xHigh)","organization":"xai","score":0.00888656313352243,"low":0.008063885461040811,"high":0.009709240806004047,"observations":3300736,"sourceRank":14,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-6 Luna (Max)","organization":"openai","score":0.008877772236971282,"low":0.008055307515136034,"high":0.00970023695880653,"observations":36929317,"sourceRank":15,"publishedAt":"2026-11-04","rank":1},{"name":"GLM-5.2 (Max)","organization":"zai","score":0.008867736420469896,"low":0.008042969765002737,"high":0.009692503075937055,"observations":8374554,"sourceRank":16,"publishedAt":"2026-11-04","rank":1},{"name":"GPT-5.6 Terra (xHigh)","organization":"openai","score":0.00875409778973335,"low":0.007904535412275999,"high":0.009603660167190702,"observations":3291847,"sourceRank":17,"publishedAt":"2026-11-04","rank":1},{"name":"Muse Spark 1.1","organization":"meta","score":0.008512184691632285,"low":0.007616512986268945,"high":0.009407856396995625,"observations":7669418,"sourceRank":18,"publishedAt":"2026-11-04","rank":1},{"name":"Muse Spark 1.3 (Max)","organization":"meta","score":0.008437635265977406,"low":0.007550209249790619,"high":0.009325061282164192,"observations":7804971,"sourceRank":19,"publishedAt":"2026-11-04","rank":1},{"name":"Claude Opus 5 (Max)","organization":"anthropic","score":0.008429781934204073,"low":0.007507095691413535,"high":0.00935246817699461,"observations":4255614,"sourceRank":20,"publishedAt":"2026-11-04","rank":1},{"name":"Claude Fable 5.1 (Max)","organization":"anthropic","score":0.0076794161975186626,"low":0.006375619146448919,"high":0.008983213248588405,"observations":2223941,"sourceRank":21,"publishedAt":"2026-11-04","rank":1},{"name":"Claude Sonnet 5.5 (Max)","organization":"anthropic","score":0.007632354061946955,"low":0.006280185099449493,"high":0.008984523024444418,"observations":1040676,"sourceRank":22,"publishedAt":"2026-11-04","rank":1},{"name":"Muse Spark 1.2 (xHigh)","organization":"meta","score":0.0072415743289937495,"low":0.005165795096351498,"high":0.009317353561636002,"observations":3315530,"sourceRank":23,"publishedAt":"2026-11-04","rank":1},{"name":"Claude Opus 5.5 (Max)","organization":"anthropic","score":0.006572867773321889,"low":0.004702545595106526,"high":0.008443189951537252,"observations":2043959,"sourceRank":24,"publishedAt":"2026-11-04","rank":1},{"name":"Claude Fable 5 (High)","organization":"anthropic","score":0.00612490514266284,"low":0.00413799074130359,"high":0.00811181954402209,"observations":3140962,"sourceRank":25,"publishedAt":"2026-11-04","rank":1},{"name":"Gemini 3.7 Flash (High)","organization":"google","score":0.005844766527140166,"low":0.004395817569293017,"high":0.007293715484987315,"observations":5594205,"sourceRank":26,"publishedAt":"2026-11-04","rank":21},{"name":"Gemini 4 Carbon (High)","organization":"google","score":0.005127644772258577,"low":0.0031273541864787215,"high":0.007127935358038432,"observations":6732336,"sourceRank":27,"publishedAt":"2026-11-04","rank":21},{"name":"Claude Sonnet 5 (High)","organization":"anthropic","score":0.005118701087418742,"low":0.003673492934262495,"high":0.006563909240574989,"observations":4781638,"sourceRank":28,"publishedAt":"2026-11-04","rank":21},{"name":"Gemini 4 Argon (High)","organization":"google","score":0.004166124522534176,"low":0.002689597656659024,"high":0.0056426513884093275,"observations":5809460,"sourceRank":29,"publishedAt":"2026-11-04","rank":23},{"name":"Gemini 3.8 Flash (High)","organization":"google","score":0.004115174279824885,"low":0.0025261060684887995,"high":0.0057042424911609695,"observations":7341646,"sourceRank":30,"publishedAt":"2026-11-04","rank":23},{"name":"DeepSeek-V4.1-Flash-20261017 (Max)","organization":"deepseek","score":0.0034828698559437044,"low":0.0024255916006196733,"high":0.004540148111267735,"observations":17216312,"sourceRank":31,"publishedAt":"2026-11-04","rank":25},{"name":"MiniMax-M3.1-Flash","organization":"minimax","score":0.002605696120994594,"low":0.0003894751980106255,"high":0.004821917043978562,"observations":2191528,"sourceRank":32,"publishedAt":"2026-11-04","rank":24},{"name":"Hy4 preview","organization":"tencent","score":0.0008978785661284405,"low":-0.0011143455628242313,"high":0.0029101026950811124,"observations":10128437,"sourceRank":33,"publishedAt":"2026-11-04","rank":29},{"name":"Qwen3.8-27B","organization":"alibaba","score":0.0007570044053721858,"low":-0.0011455873089718416,"high":0.0026595961197162133,"observations":7754962,"sourceRank":34,"publishedAt":"2026-11-04","rank":30},{"name":"Gemini 3.1 Pro Preview","organization":"google","score":-0.0002536900357057448,"low":-0.006392869636950765,"high":0.005885489565539276,"observations":2736062,"sourceRank":35,"publishedAt":"2026-11-04","rank":23},{"name":"MiMo-V2.6-Pro","organization":"xiaomi","score":-0.0005668122775333018,"low":-0.002962998642937544,"high":0.0018293740878709405,"observations":4554607,"sourceRank":36,"publishedAt":"2026-11-04","rank":32},{"name":"Muse Wonder Full (Max)","organization":"meta","score":-0.0015256276953365955,"low":-0.013960982766840623,"high":0.010909727376167432,"observations":5071624,"sourceRank":37,"publishedAt":"2026-11-04","rank":1},{"name":"Mistral Large 4","organization":"mistral","score":-0.0016103030452185862,"low":-0.012065901884917532,"high":0.00884529579448036,"observations":386267,"sourceRank":38,"publishedAt":"2026-11-04","rank":1},{"name":"Qwen3.7-Max","organization":"alibaba","score":-0.002866240270407161,"low":-0.007023129669838171,"high":0.0012906491290238494,"observations":2107566,"sourceRank":39,"publishedAt":"2026-11-04","rank":32},{"name":"Hy4","organization":"tencent","score":-0.0037060989909747377,"low":-0.010677178175228376,"high":0.0032649801932788996,"observations":1314198,"sourceRank":40,"publishedAt":"2026-11-04","rank":28},{"name":"Step 5 Preview","organization":"stepfun","score":-0.004842784070866912,"low":-0.00859761618136341,"high":-0.0010879519603704162,"observations":2427836,"sourceRank":41,"publishedAt":"2026-11-04","rank":33},{"name":"Qwen3.8-Max","organization":"alibaba","score":-0.007335158415322951,"low":-0.009506294213055074,"high":-0.005164022617590828,"observations":7343430,"sourceRank":42,"publishedAt":"2026-11-04","rank":36},{"name":"Muse Wonder (Max)","organization":"meta","score":-0.007338526410606505,"low":-0.02343565336915256,"high":0.00875860054793955,"observations":1745681,"sourceRank":43,"publishedAt":"2026-11-04","rank":1},{"name":"Qwen3.8-Flash-Next","organization":"alibaba","score":-0.008963140952168658,"low":-0.010838281024754023,"high":-0.007088000879583293,"observations":15320065,"sourceRank":44,"publishedAt":"2026-11-04","rank":38},{"name":"MiMo-V2.5-Pro","organization":"xiaomi","score":-0.009511148497054478,"low":-0.013192558221566144,"high":-0.005829738772542811,"observations":2319225,"sourceRank":45,"publishedAt":"2026-11-04","rank":36},{"name":"Beam 1.0 501B A23B","organization":"reflection","score":-0.010393626022462854,"low":-0.01925688944764077,"high":-0.0015303625972849372,"observations":555707,"sourceRank":46,"publishedAt":"2026-11-04","rank":35},{"name":"MiniMax-M3","organization":"minimax","score":-0.013539751793885557,"low":-0.03467998058287114,"high":0.007600476995100025,"observations":3464873,"sourceRank":47,"publishedAt":"2026-11-04","rank":19},{"name":"Claude Haiku 5.5 (High)","organization":"anthropic","score":-0.014678555003932559,"low":-0.021542545208415365,"high":-0.007814564799449754,"observations":3371105,"sourceRank":48,"publishedAt":"2026-11-04","rank":38},{"name":"Nemotron 3 Ultra","organization":"nvidia","score":-0.017225615481315515,"low":-0.04139884922761437,"high":0.0069476182649833435,"observations":209335,"sourceRank":49,"publishedAt":"2026-11-04","rank":21},{"name":"Spark-X2.6","organization":"synlan","score":-0.019552848687339974,"low":-0.03215539331226018,"high":-0.0069503040624197664,"observations":856437,"sourceRank":50,"publishedAt":"2026-11-04","rank":37},{"name":"Inkling","organization":"thinky","score":-0.020014304210391742,"low":-0.034228755247579803,"high":-0.005799853173203678,"observations":1645936,"sourceRank":51,"publishedAt":"2026-11-04","rank":36},{"name":"Qwen3.7-Plus","organization":"alibaba","score":-0.021771644534865686,"low":-0.03435713450702523,"high":-0.00918615456270614,"observations":1177145,"sourceRank":52,"publishedAt":"2026-11-04","rank":39},{"name":"Inkling-Small","organization":"thinky","score":-0.027930110325897317,"low":-0.043248907241297325,"high":-0.012611313410497307,"observations":499930,"sourceRank":53,"publishedAt":"2026-11-04","rank":43},{"name":"Solar Pro 4","organization":"upstage","score":-0.03742620205242986,"low":-0.055086900315680504,"high":-0.019765503789179212,"observations":488500,"sourceRank":54,"publishedAt":"2026-11-04","rank":46},{"name":"MiMo-V2.6-Flash","organization":"xiaomi","score":-0.041881118768912984,"low":-0.06361556220994041,"high":-0.020146675327885555,"observations":3121990,"sourceRank":55,"publishedAt":"2026-11-04","rank":46},{"name":"Mistral Medium 3.5","organization":"mistral","score":-0.04517795203449282,"low":-0.06514422942612771,"high":-0.025211674642857923,"observations":599909,"sourceRank":56,"publishedAt":"2026-11-04","rank":48},{"name":"Hy3","organization":"tencent","score":-0.04696924288973747,"low":-0.06996986272006377,"high":-0.023968623059411182,"observations":1149676,"sourceRank":57,"publishedAt":"2026-11-04","rank":48},{"name":"NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)","organization":"nvidia","score":-0.07292189639386493,"low":-0.08688080053669185,"high":-0.058962992251038,"observations":1336586,"sourceRank":58,"publishedAt":"2026-11-04","rank":55}],"source":{"id":"arena","publisher":"Arena (lmarena-ai)","href":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","licence":"CC-BY-4.0","licenceUrl":"https://creativecommons.org/licenses/by/4.0/","revision":"8b6deb69ace01ee8dcda9e46eba7a5e8a808ec53","fetchedAt":"2026-11-05T06:35:05.520Z","evidence":[{"file":"a7a159f557a1e594a5d41218fe8d53c63c63daa39b655e51d6fc5b482a17d059.json","sha256":"a7a159f557a1e594a5d41218fe8d53c63c63daa39b655e51d6fc5b482a17d059","url":"https://huggingface.co/api/datasets/lmarena-ai/leaderboard-dataset","revision":null,"role":"metadata"},{"file":"88a46d472f35274fb6b7a852c3d16c8ee8007d51b37f290161537bec26895ee4.txt","sha256":"88a46d472f35274fb6b7a852c3d16c8ee8007d51b37f290161537bec26895ee4","url":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/raw/8b6deb69ace01ee8dcda9e46eba7a5e8a808ec53/README.md","revision":"8b6deb69ace01ee8dcda9e46eba7a5e8a808ec53","role":"licence"},{"file":"bb1b9d40eeca30c1ab54a3b90717a21e2daa4d34bb356054851b3bb7f3390fe6.parquet","sha256":"bb1b9d40eeca30c1ab54a3b90717a21e2daa4d34bb356054851b3bb7f3390fe6","url":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/resolve/8b6deb69ace01ee8dcda9e46eba7a5e8a808ec53/agent_tool_hallucination/latest-00000-of-00001.parquet","revision":"8b6deb69ace01ee8dcda9e46eba7a5e8a808ec53","role":"rows"}]},"status":"fresh"}