Benchmarks

综合能力

GDPval-AA v2.1

真实职场任务:能否交付可用的文档、表格、幻灯片和图示

它测什么

GDPval-AA v2.1 是 Artificial Analysis 基于 OpenAI GDPval 数据集进行的独立评测。它使用公开 gold 集中的 220 个任务,覆盖 9 个行业、44 种职业,要求模型提交实际工作成果,而不只是回答问题。

比如,官方示例会让模型给一支巡演乐队制作一页舞台布置 PDF:把歌手、鼓手、音箱和话筒放对位置,再列清楚输入、输出线路。能说出“舞台图应该怎么画”还不够,模型需要真正做出文件,内容和排版都要满足需求。

所有模型通过 Stirrup Agent 框架工作,可以使用命令行、文件处理、网页搜索与浏览等工具,每个任务最多执行 250 轮。评估时隐藏模型身份,把同一任务的两份成果交给从三模型评审团中抽取的 LLM 裁判比较,再将这些对比结果汇总成 Elo 评分。

怎么看分数

Elo 越高,表示交付成果在这套任务的两两比较中更受评审偏好;DeepSeek V4.1 Flash (Max) 固定为 1,600 分作为锚点,其他模型都相对它计分。比如 A 得 1,500 分、B 得 1,200 分,只能说明 A 在这一评测中的相对表现更强,不表示 A 的正确率是 150%,也不表示效率高 25%。Elo 可以低于 0,表中的 ± 数值是官方公布的 95% 置信区间,分数接近时不要把名次差当成确定差距。本页保留 2026 年 11 月 1 日核对到的全部 308 条成绩和官方模型配置;标有 Fallback 的条目包含备用模型机制,不能当成单一模型独立完成的结果。v2.1 沿用 v2 的任务、工具和评审流程,但更换了锚点并改用 Crowd-BT 模型拟合评分,分数整体平移,因此不能与 v2 快照混用。榜单会随新增比较而变化,也不能直接与 v1、模型发布时的旧快照或 OpenAI 自己公布的 GDPval 胜率混用;自 Intelligence Index v5.0 起,它也不再计入该指数,是一项单独的评测。

来源Artificial Analysis

查看 GDPval-AA v2.1 官方评测方法

模型表现

核对于 308 条成绩

GDPval-AA v2.1 — 模型表现。选择列标题可排序。
Claude Opus 5.5 (Max, Default Fallback)AnthropicAgent · Stirrup
Stirrup1866± 26
Kimi K3.1 ProKimiAgent · Stirrup
Stirrup1858± 28
Muse Wonder Full (Max)MetaAgent · Stirrup
Stirrup1857± 25
Claude Sonnet 5.5 (Max, Default Fallback)AnthropicAgent · Stirrup
Stirrup1838± 23
Claude Opus 5.5 (Xhigh, Default Fallback)AnthropicAgent · Stirrup
Stirrup1835± 23
Muse Wonder (Max)MetaAgent · Stirrup
Stirrup1770± 23
Claude Fable 5.1 (Max, Default Fallback)AnthropicAgent · Stirrup
Stirrup1763± 19
DeepSeek V4.1 Flash 1017 (Max)DeepSeekAgent · Stirrup
Stirrup1746± 21
Claude Fable 5.1 (Xhigh, Default Fallback)AnthropicAgent · Stirrup
Stirrup1740± 19
Claude Sonnet 5.5 (Xhigh, Default Fallback)AnthropicAgent · Stirrup
Stirrup1735± 23
Gemini 4 Carbon (High)GoogleAgent · Stirrup
Stirrup1733± 24
Claude Opus 5 (Max)AnthropicAgent · Stirrup
Stirrup1730± 19
Grok 4.7 (Xhigh)SpaceXAIAgent · Stirrup
Stirrup1718± 21
Grok 4.7 (High)SpaceXAIAgent · Stirrup
Stirrup1716± 19
Claude Opus 5.5 (High, Default Fallback)AnthropicAgent · Stirrup
Stirrup1714± 21
Claude Opus 5 (Xhigh)AnthropicAgent · Stirrup
Stirrup1696± 18
MiMo-V2.6-ProXiaomiAgent · Stirrup
Stirrup1689± 21
Muse Spark 1.3 (Max)MetaAgent · Stirrup
Stirrup1681± 19
Qwen3.8 Max (0902)AlibabaAgent · Stirrup
Stirrup1674± 20
GLM-5.3 (Max)Z AIAgent · Stirrup
Stirrup1657± 19
Grok 4.6 (Xhigh)SpaceXAIAgent · Stirrup
Stirrup1654± 19
Ling 3.1 Flash 1019InclusionAIAgent · Stirrup
Stirrup1653± 25
GLM-5.3-FlashZ AIAgent · Stirrup
Stirrup1649± 17
Claude Fable 5.1 (High, Default Fallback)AnthropicAgent · Stirrup
Stirrup1644± 18
Qwen3.8-Flash-NextAlibabaAgent · Stirrup
Stirrup1636± 18
Muse Spark 1.3 (Xhigh)MetaAgent · Stirrup
Stirrup1632± 18
Claude Haiku 5.5 (Max)AnthropicAgent · Stirrup
Stirrup1631± 22
Gemini 4 Argon (High)GoogleAgent · Stirrup
Stirrup1631± 20
Ling 3.1 FlashInclusionAIAgent · Stirrup
Stirrup1626± 21
Grok 4.6 (High)SpaceXAIAgent · Stirrup
Stirrup1626± 16
Grok 4.6 (Medium)SpaceXAIAgent · Stirrup
Stirrup1620± 18
Claude Fable 5 (Max, Opus 4.8 Fallback)AnthropicAgent · Stirrup
Stirrup1618± 16
GPT-5.6 Sol (Max)OpenAIAgent · Stirrup
Stirrup1618± 16
Qwen3.8 2.4T A95BAlibabaAgent · Stirrup
Stirrup1616± 19
MiMo-V2.6-FlashXiaomiAgent · Stirrup
Stirrup1614± 21
Qwen3.8 MaxAlibabaAgent · Stirrup
Stirrup1611± 17
GPT-6.1 Sol (Max)OpenAIAgent · Stirrup
Stirrup1609± 22
Claude Opus 5 (High)AnthropicAgent · Stirrup
Stirrup1604± 18
DeepSeek V4.1 Flash (Max)DeepSeekAgent · Stirrup
Stirrup16001
Step 5 PreviewStepFunAgent · Stirrup
Stirrup1594± 23
Claude Opus 5.5 (Medium, Default Fallback)AnthropicAgent · Stirrup
Stirrup1593± 20
Grok 4.7 (Low)SpaceXAIAgent · Stirrup
Stirrup1591± 22
GPT-6 Astra (Max)OpenAIAgent · Stirrup
Stirrup1584± 19
GPT-5.6 Sol (Xhigh)OpenAIAgent · Stirrup
Stirrup1578± 16
GPT-6 Astra (Xhigh)OpenAIAgent · Stirrup
Stirrup1569± 19
Claude Sonnet 5.5 (High, Default Fallback)AnthropicAgent · Stirrup
Stirrup1557± 22
GPT-6.1 Sol (Xhigh)OpenAIAgent · Stirrup
Stirrup1555± 21
Claude Fable 5.1 (Medium, Default Fallback)AnthropicAgent · Stirrup
Stirrup1554± 18
DeepSeek V4 Flash Vision (Max)DeepSeekAgent · Stirrup
Stirrup1554± 17
Kimi K3 (Max)KimiAgent · Stirrup
Stirrup1543± 18
GPT-6 Astra (High)OpenAIAgent · Stirrup
Stirrup1536± 19
Claude Haiku 5.5 (Xhigh)AnthropicAgent · Stirrup
Stirrup1522± 22
GPT-6.1 Sol (High)OpenAIAgent · Stirrup
Stirrup1520± 21
GPT-6 Sol (Max)OpenAIAgent · Stirrup
Stirrup1516± 22
GPT-5.6 Sol (High)OpenAIAgent · Stirrup
Stirrup1512± 16
Claude Opus 5 (Medium)AnthropicAgent · Stirrup
Stirrup1499± 18
GPT-6 Astra (Medium)OpenAIAgent · Stirrup
Stirrup1498± 18
Muse Spark 1.2 (Xhigh)MetaAgent · Stirrup
Stirrup1480± 19
Claude Fable 5.1 (Low, Default Fallback)AnthropicAgent · Stirrup
Stirrup1477± 18
GPT-6.1 Terra (Max)OpenAIAgent · Stirrup
Stirrup1476± 23
Claude Sonnet 5 (Max)AnthropicAgent · Stirrup
Stirrup1470± 16
GPT-5.6 Luna (Max)OpenAIAgent · Stirrup
Stirrup1468± 16
GPT-6 Sol (Xhigh)OpenAIAgent · Stirrup
Stirrup1466± 22
MiniMax-M3.1-FlashMiniMaxAgent · Stirrup
Stirrup1465± 23
Claude Opus 4.8 (Max)AnthropicAgent · Stirrup
Stirrup1462± 15
GPT-5.6 Terra (Max)OpenAIAgent · Stirrup
Stirrup1461± 16
GPT-6.1 Sol (Medium)OpenAIAgent · Stirrup
Stirrup1459± 21
DeepSeek V4 Pro 0813 (Max)DeepSeekAgent · Stirrup
Stirrup1457± 19
GPT-5.6 Terra (Xhigh)OpenAIAgent · Stirrup
Stirrup1455± 16
Gemini 3.8 Flash (High)GoogleAgent · Stirrup
Stirrup1445± 19
GPT-6 Luna (Max)OpenAIAgent · Stirrup
Stirrup1444± 20
DeepSeek V4 Flash 0731 (Max)DeepSeekAgent · Stirrup
Stirrup1441± 19
Mistral Large 4 PreviewMistralAgent · Stirrup
Stirrup1433± 23
GPT-5.6 Sol (Medium)OpenAIAgent · Stirrup
Stirrup1432± 16
Gemini 3.8 Flash (Medium)GoogleAgent · Stirrup
Stirrup1431± 20
Qwen3.8 27B (Xhigh)AlibabaAgent · Stirrup
Stirrup1427± 20
Claude Haiku 5.5 (High)AnthropicAgent · Stirrup
Stirrup1424± 19
Grok 4.6 (Low)SpaceXAIAgent · Stirrup
Stirrup1418± 19
GPT-6.1 Terra (Xhigh)OpenAIAgent · Stirrup
Stirrup1415± 23
GPT-5.6 Luna (Xhigh)OpenAIAgent · Stirrup
Stirrup1408± 17
GPT-6 Sol (High)OpenAIAgent · Stirrup
Stirrup1404± 21
GPT-6 Astra (Low)OpenAIAgent · Stirrup
Stirrup1403± 20
Gemini 3.7 Flash (High)GoogleAgent · Stirrup
Stirrup1403± 19
Muse Spark Open (Xhigh)MetaAgent · Stirrup
Stirrup1401± 19
Grok 4.5 (High)SpaceXAIAgent · Stirrup
Stirrup1389± 19
Qwen3.8 27B (Medium)AlibabaAgent · Stirrup
Stirrup1389± 18
GPT-5.6 Terra (High)OpenAIAgent · Stirrup
Stirrup1387± 17
GLM-5.2 (Max)Z AIAgent · Stirrup
Stirrup1381± 16
GPT-6 Luna (Xhigh)OpenAIAgent · Stirrup
Stirrup1373± 21
Qwen3.8 27B (Low)AlibabaAgent · Stirrup
Stirrup1370± 19
Gemini 3.7 Flash (Medium)GoogleAgent · Stirrup
Stirrup1367± 20
Claude Sonnet 5 (Xhigh)AnthropicAgent · Stirrup
Stirrup1366± 17
K2 Horizon 375B A23BMBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup1365± 19
Claude Opus 4.7 (Max)AnthropicAgent · Stirrup
Stirrup1362± 16
GPT-5.5 (Xhigh)OpenAIAgent · Stirrup
Stirrup1359± 16
GPT-6.1 Terra (High)OpenAIAgent · Stirrup
Stirrup1353± 23
GPT-6 Luna (High)OpenAIAgent · Stirrup
Stirrup1352± 21
GPT-6 Sol (Medium)OpenAIAgent · Stirrup
Stirrup1352± 22
DeepSeek V4.1 Flash (Non-reasoning)DeepSeekAgent · Stirrup
Stirrup1349± 20
GPT-5.6 Luna (High)OpenAIAgent · Stirrup
Stirrup1339± 17
GPT-5.5 (High)OpenAIAgent · Stirrup
Stirrup1336± 16
Agnes 2.5 Pro BetaSapiens AIAgent · Stirrup
Stirrup1335± 20
JT-4.1 Flash 236B A21B (Reasoning)China MobileAgent · Stirrup
Stirrup1335± 23
GLM-5.3 (Low)Z AIAgent · Stirrup
Stirrup1333± 20
Claude Sonnet 5.5 (Medium, Default Fallback)AnthropicAgent · Stirrup
Stirrup1329± 21
Gemini 3.7 Flash (Low)GoogleAgent · Stirrup
Stirrup1327± 20
GPT-6.1 Terra (Medium)OpenAIAgent · Stirrup
Stirrup1325± 22
GPT-6.1 Sol (Low)OpenAIAgent · Stirrup
Stirrup1322± 21
Gemini 3.8 Flash (Low)GoogleAgent · Stirrup
Stirrup1316± 17
GPT-5.6 Sol (Low)OpenAIAgent · Stirrup
Stirrup1313± 17
Claude Opus 5 (Low)AnthropicAgent · Stirrup
Stirrup1311± 20
Gemini 3.6 Flash (High)GoogleAgent · Stirrup
Stirrup1295± 19
GPT-6.1 Terra (Non-reasoning)OpenAIAgent · Stirrup
Stirrup1286± 23
Claude Haiku 5.5 (Medium)AnthropicAgent · Stirrup
Stirrup1283± 20
GPT-5.6 Terra (Medium)OpenAIAgent · Stirrup
Stirrup1274± 17
GPT-6 Luna (Medium)OpenAIAgent · Stirrup
Stirrup1270± 21
Claude Sonnet 5 (High)AnthropicAgent · Stirrup
Stirrup1263± 17
GPT-5.4 (Xhigh)OpenAIAgent · Stirrup
Stirrup1256± 16
GPT-6 Sol (Non-reasoning)OpenAIAgent · Stirrup
Stirrup1254± 21
GLM-5.2 (Non-reasoning)Z AIAgent · Stirrup
Stirrup1252± 21
MiniMax-M3MiniMaxAgent · Stirrup
Stirrup1250± 16
GPT-5.6 Sol (Non-reasoning)OpenAIAgent · Stirrup
Stirrup1250± 18
GPT-5.5 (Medium)OpenAIAgent · Stirrup
Stirrup1243± 18
Claude Sonnet 4.6 (Max)AnthropicAgent · Stirrup
Stirrup1241± 16
Claude Opus 5.5 (Low, Default Fallback)AnthropicAgent · Stirrup
Stirrup1240± 21
Claude Sonnet 5 (Non-reasoning, High)AnthropicAgent · Stirrup
Stirrup1221± 19
Muse Spark 1.1 (Xhigh)MetaAgent · Stirrup
Stirrup1217± 20
Gemini 3.5 Flash (High)GoogleAgent · Stirrup
Stirrup1212± 17
GPT-6.1 Terra (Low)OpenAIAgent · Stirrup
Stirrup1204± 23
GPT-6 Sol (Low)OpenAIAgent · Stirrup
Stirrup1201± 22
K2 Horizon MoVA 36B A4BMBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup1197± 21
Apodex 1.1ApodexAgent · Stirrup
Stirrup1196± 20
Quasar 438B (Max, Based on GLM-5.2)Multiverse ComputingAgent · Stirrup
Stirrup1190± 20
Claude Sonnet 5.5 (Low, Default Fallback)AnthropicAgent · Stirrup
Stirrup1182± 22
Beam 1.0ReflectionAgent · Stirrup
Stirrup1177± 19
Ling-3.0-flash-VLInclusionAIAgent · Stirrup
Stirrup1169± 21
Claude Sonnet 5 (Medium)AnthropicAgent · Stirrup
Stirrup1167± 18
DeepSeek V4 Pro 0424 (Max)DeepSeekAgent · Stirrup
Stirrup1166± 17
Nemotron 3.5 Super VLNVIDIAAgent · Stirrup
Stirrup1162± 19
DeepSeek V4 Pro 0424 (High)DeepSeekAgent · Stirrup
Stirrup1160± 22
GPT-5.6 Luna (Medium)OpenAIAgent · Stirrup
Stirrup1138± 19
Kimi K3 (Low)KimiAgent · Stirrup
Stirrup1136± 22
Qwen3.7 MaxAlibabaAgent · Stirrup
Stirrup1135± 17
Inkling SmallThinking MachinesAgent · Stirrup
Stirrup1133± 22
Claude Haiku 5.5 (Low)AnthropicAgent · Stirrup
Stirrup1130± 22
MiMo-V2.5-Pro (Reasoning)XiaomiAgent · Stirrup
Stirrup1129± 17
JT-4.1 Flash 236B A21B (Non-reasoning)China MobileAgent · Stirrup
Stirrup1128± 21
Motif 3Motif TechnologiesAgent · Stirrup
Stirrup1127± 22
GLM-5.1 (Reasoning)Z AIAgent · Stirrup
Stirrup1125± 17
GPT-5.6 Terra (Low)OpenAIAgent · Stirrup
Stirrup1117± 18
Nex-N2-Pro (Based on Qwen3.5-397B-A17B)Nex AGIAgent · Stirrup
Stirrup1109± 19
G9v3-39A5BAI9StarsAgent · Stirrup
Stirrup1109± 23
Motif 3 (Beta)Motif TechnologiesAgent · Stirrup
Stirrup1109± 22
Ling-3.0-flash-FinInclusionAIAgent · Stirrup
Stirrup1108± 22
Solar Pro 4UpstageAgent · Stirrup
Stirrup1106± 20
GPT-5.6 Terra (Non-reasoning)OpenAIAgent · Stirrup
Stirrup1105± 19
Solar Mini 4UpstageAgent · Stirrup
Stirrup1092± 23
Qwen3.8 27B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup1082± 20
Inkling (Xhigh)Thinking MachinesAgent · Stirrup
Stirrup1080± 22
Claude Sonnet 5 (Low)AnthropicAgent · Stirrup
Stirrup1076± 18
Grok Build 0.1 0616SpaceXAIAgent · Stirrup
Stirrup1073± 18
Hy3TencentAgent · Stirrup
Stirrup1071± 22
DeepSeek V4 Pro 0813 (Non-reasoning)DeepSeekAgent · Stirrup
Stirrup1066± 22
GPT-5.5 (Low)OpenAIAgent · Stirrup
Stirrup1053± 19
GPT-6 Luna (Non-reasoning)OpenAIAgent · Stirrup
Stirrup1052± 22
Kimi K2.6 (Reasoning)KimiAgent · Stirrup
Stirrup1047± 17
Kimi K2.7 CodeKimiAgent · Stirrup
Stirrup1046± 17
DeepSeek V4 Flash 0420 (Max)DeepSeekAgent · Stirrup
Stirrup1046± 18
GPT-6 Luna (Low)OpenAIAgent · Stirrup
Stirrup1040± 22
K2 Horizon 7BMBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup1024± 22
GPT-5.4 mini (Xhigh)OpenAIAgent · Stirrup
Stirrup1022± 17
GLM-4.7 (Reasoning)Z AIAgent · Stirrup
Stirrup1021± 20
Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIAAgent · Stirrup
Stirrup1021± 17
MiniMax-M2.7MiniMaxAgent · Stirrup
Stirrup1016± 17
DeepSeek V4 Flash 0420 (High)DeepSeekAgent · Stirrup
Stirrup1009± 23
Muse SparkMetaAgent · Stirrup
Stirrup1006± 17
MiMo-V2.5XiaomiAgent · Stirrup
Stirrup1005± 22
GPT-5.6 Luna (Low)OpenAIAgent · Stirrup
Stirrup1002± 19
Qwen3.6 PlusAlibabaAgent · Stirrup
Stirrup997± 17
Gemini 3.5 Flash-LiteGoogleAgent · Stirrup
Stirrup995± 22
Qwen3.6 27B (Reasoning)AlibabaAgent · Stirrup
Stirrup993± 17
Nemotron 3.5 Bolt 74B-A9BNVIDIAAgent · Stirrup
Stirrup979± 19
Solar Open2 250BUpstageAgent · Stirrup
Stirrup978± 24
GPT-5.5 (Non-reasoning)OpenAIAgent · Stirrup
Stirrup978± 17
Qwen3.6 27B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup969± 21
A.X-K2SK TelecomAgent · Stirrup
Stirrup968± 23
Ling 3.0 FlashInclusionAIAgent · Stirrup
Stirrup957± 23
GPT-5.4 nano (Xhigh)OpenAIAgent · Stirrup
Stirrup955± 17
Grok 4.3 (Non-reasoning)SpaceXAIAgent · Stirrup
Stirrup949± 18
Grok 4.3 (High)SpaceXAIAgent · Stirrup
Stirrup946± 18
GPT-5.6 Luna (Non-reasoning)OpenAIAgent · Stirrup
Stirrup928± 20
GPT-5 (High)OpenAIAgent · Stirrup
Stirrup925± 21
Claude 4.5 Sonnet (Reasoning)AnthropicAgent · Stirrup
Stirrup915± 21
Qwen3.6 35B A3B (Reasoning)AlibabaAgent · Stirrup
Stirrup899± 18
K2 Horizon 3.7BMBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup889± 24
LongCat 2.0LongCatAgent · Stirrup
Stirrup877± 21
Step 3.7 FlashStepFunAgent · Stirrup
Stirrup864± 18
Qwen3.6 35B A3B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup854± 24
Kimi K2.5 (Reasoning)KimiAgent · Stirrup
Stirrup847± 23
GPT-5.1 (High)OpenAIAgent · Stirrup
Stirrup834± 22
Qwen3.5 122B A10B (Reasoning)AlibabaAgent · Stirrup
Stirrup822± 18
K-EXAONE 2.0 0803LG AI ResearchAgent · Stirrup
Stirrup816± 25
Qwen3.5 397B A17B (Reasoning)AlibabaAgent · Stirrup
Stirrup801± 19
Gemini 3.1 Pro PreviewGoogleAgent · Stirrup
Stirrup799± 19
Muse Glimmer (High)MetaAgent · Stirrup
Stirrup792± 23
Qwen3.7 PlusAlibabaAgent · Stirrup
Stirrup775± 19
GPT-5 mini (High)OpenAIAgent · Stirrup
Stirrup773± 22
Mistral Medium 3.5MistralAgent · Stirrup
Stirrup769± 19
GLM-4.6 (Reasoning)Z AIAgent · Stirrup
Stirrup764± 22
Ring-2.6-1TInclusionAIAgent · Stirrup
Stirrup753± 19
KAT Coder Pro V2KwaiKATAgent · Stirrup
Stirrup742± 23
Claude 4.5 Haiku (Reasoning)AnthropicAgent · Stirrup
Stirrup740± 19
KAT-Coder-Pro V1KwaiKATAgent · Stirrup
Stirrup738± 21
DeepSeek V3.1 Terminus (Reasoning)DeepSeekAgent · Stirrup
Stirrup718± 24
Qwen3.5 122B A10B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup716± 23
MiniCPM5-2BOpenBMBAgent · Stirrup
Stirrup715± 25
DeepSeek V3.2 (Reasoning)DeepSeekAgent · Stirrup
Stirrup699± 25
Claude 4 Sonnet (Reasoning)AnthropicAgent · Stirrup
Stirrup698± 22
Nemotron 3.5 LightningNVIDIAAgent · Stirrup
Stirrup646± 24
MiMo-V2-Flash (Non-reasoning)XiaomiAgent · Stirrup
Stirrup636± 24
Gemma 4 31B (Reasoning)GoogleAgent · Stirrup
Stirrup627± 20
Qwen3.5 35B A3B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup616± 25
gpt-oss-120b (High)OpenAIAgent · Stirrup
Stirrup616± 19
GPT-5.4 mini (Non-reasoning)OpenAIAgent · Stirrup
Stirrup597± 20
Granite 4.2 30BIBMAgent · Stirrup
Stirrup583± 23
Gemma 4 26B A4B (Reasoning)GoogleAgent · Stirrup
Stirrup573± 19
Ling 3.0 TinyInclusionAIAgent · Stirrup
Stirrup568± 26
Gemma 4 31B (Non-reasoning)GoogleAgent · Stirrup
Stirrup555± 23
Devstral 2MistralAgent · Stirrup
Stirrup552± 21
Devstral Small 2MistralAgent · Stirrup
Stirrup535± 20
Command A+CohereAgent · Stirrup
Stirrup522± 22
Qwen3 Coder NextAlibabaAgent · Stirrup
Stirrup521± 23
GPT-5.5 Instant (June 2026)OpenAIAgent · Stirrup
Stirrup516± 21
Mercury 2InceptionAgent · Stirrup
Stirrup502± 23
Granite 4.2 8BIBMAgent · Stirrup
Stirrup499± 24
Nemotron 3 Super 120B A12B (Reasoning)NVIDIAAgent · Stirrup
Stirrup497± 19
Mercury 2.5InceptionAgent · Stirrup
Stirrup482± 25
Nova 2.0 Pro Preview (Medium)AmazonAgent · Stirrup
Stirrup477± 19
EXAONE 4.5 33B (Reasoning)LG AI ResearchAgent · Stirrup
Stirrup474± 24
Gemini 2.5 ProGoogleAgent · Stirrup
Stirrup463± 21
HyperNova 60B 2605 (High, Based on gpt-oss-120b)Multiverse ComputingAgent · Stirrup
Stirrup449± 23
Nova 2.0 Pro Preview (Low)AmazonAgent · Stirrup
Stirrup444± 20
Gemini 3.1 Flash-LiteGoogleAgent · Stirrup
Stirrup438± 19
Gemma 4 12B (Reasoning)GoogleAgent · Stirrup
Stirrup435± 24
Mistral Large 3MistralAgent · Stirrup
Stirrup428± 20
Qwen3.5 9B (Reasoning)AlibabaAgent · Stirrup
Stirrup427± 22
Mistral Medium 3.1MistralAgent · Stirrup
Stirrup393± 21
Mistral Small 3.1MistralAgent · Stirrup
Stirrup384± 21
K-EXAONE (Reasoning)LG AI ResearchAgent · Stirrup
Stirrup375± 24
Nova 2.0 Lite (High)AmazonAgent · Stirrup
Stirrup374± 20
Mistral Small 4 (Reasoning)MistralAgent · Stirrup
Stirrup370± 22
gpt-oss-20b (High)OpenAIAgent · Stirrup
Stirrup347± 19
Trinity Large ThinkingArcee AIAgent · Stirrup
Stirrup344± 24
Nova 2.0 Pro Preview (Non-reasoning)AmazonAgent · Stirrup
Stirrup339± 20
DiffusionGemma 26B A4BGoogleAgent · Stirrup
Stirrup329± 21
Ling 2.6 FlashInclusionAIAgent · Stirrup
Stirrup326± 23
Qwen3 235B A22B 2507 (Reasoning)AlibabaAgent · Stirrup
Stirrup321± 22
North Mini CodeCohereAgent · Stirrup
Stirrup319± 22
Celeris-1CelerisAgent · Stirrup
Stirrup309± 24
DeepSeek R1 (Jan '25)DeepSeekAgent · Stirrup
Stirrup302± 22
GPT-4.1 miniOpenAIAgent · Stirrup
Stirrup277± 22
Nemotron Cascade 2 30B A3BNVIDIAAgent · Stirrup
Stirrup272± 24
Solar Pro 3UpstageAgent · Stirrup
Stirrup272± 19
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIAAgent · Stirrup
Stirrup262± 20
Ministral 3 14BMistralAgent · Stirrup
Stirrup255± 20
o3-mini (High)OpenAIAgent · Stirrup
Stirrup238± 23
Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIAAgent · Stirrup
Stirrup237± 24
Ministral 3 8BMistralAgent · Stirrup
Stirrup221± 20
Claude 3.5 HaikuAnthropicAgent · Stirrup
Stirrup218± 21
Granite 4.2 3BIBMAgent · Stirrup
Stirrup209± 25
Granite 4.1 30BIBMAgent · Stirrup
Stirrup195± 20
gpt-oss-120b (Low)OpenAIAgent · Stirrup
Stirrup179± 25
Magistral Medium 1.2MistralAgent · Stirrup
Stirrup176± 21
K2 Think V2MBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup137± 22
Qwen3 Next 80B A3B (Reasoning)AlibabaAgent · Stirrup
Stirrup136± 23
DeepSeek V3 0324DeepSeekAgent · Stirrup
Stirrup77± 22
Qwen3 30B A3B 2507 (Reasoning)AlibabaAgent · Stirrup
Stirrup75± 23
Qwen3 32B (Reasoning)AlibabaAgent · Stirrup
Stirrup41± 21
Ministral 3 3BMistralAgent · Stirrup
Stirrup37± 21
Magistral Small 1.2MistralAgent · Stirrup
Stirrup12± 22
LFM2.5-2.6BLiquid AIAgent · Stirrup
Stirrup-1± 26
GPT-4o (Nov '24)OpenAIAgent · Stirrup
Stirrup-9± 29
GPT-4OpenAIAgent · Stirrup
Stirrup-13± 22
GPT-4o miniOpenAIAgent · Stirrup
Stirrup-15± 23
Qwen3 14B (Reasoning)AlibabaAgent · Stirrup
Stirrup-17± 22
DeepSeek V3 (Dec '24)DeepSeekAgent · Stirrup
Stirrup-21± 23
Gemma 4 E4B (Reasoning)GoogleAgent · Stirrup
Stirrup-29± 25
Qwen3.5 2B (Reasoning)AlibabaAgent · Stirrup
Stirrup-39± 24
Qwen3 8B (Reasoning)AlibabaAgent · Stirrup
Stirrup-41± 23
NVIDIA Nemotron 3 Nano 4BNVIDIAAgent · Stirrup
Stirrup-51± 24
Granite 4.1 3BIBMAgent · Stirrup
Stirrup-138± 24
Llama 4 ScoutMetaAgent · Stirrup
Stirrup-156± 21
Llama 3.3 Instruct 70BMetaAgent · Stirrup
Stirrup-166± 22
Gemma 4 E2B (Reasoning)GoogleAgent · Stirrup
Stirrup-181± 24
Mistral Small 3.2MistralAgent · Stirrup
Stirrup-188± 23
GPT-4.1 nanoOpenAIAgent · Stirrup
Stirrup-206± 22
Llama 4 MaverickMetaAgent · Stirrup
Stirrup-268± 20
NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIAAgent · Stirrup
Stirrup-349± 20
Qwen3.5 2B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup-359± 21
MiniCPM-V 4.6 1.3BOpenBMBAgent · Stirrup
Stirrup-362± 20
Qwen3.5 0.8B (Non-reasoning)AlibabaAgent · Stirrup
Stirrup-363± 21
Llama 3.1 Instruct 8BMetaAgent · Stirrup
Stirrup-380± 20
Qwen3.5 0.8B (Reasoning)AlibabaAgent · Stirrup
Stirrup-381± 21
Nanbeige4.1-3BNanbeigeAgent · Stirrup
Stirrup-397± 20
Gemma 3 27B InstructGoogleAgent · Stirrup
Stirrup-401± 20
Gemma 3 12B InstructGoogleAgent · Stirrup
Stirrup-401± 20
Phi-4 Mini InstructMicrosoftAgent · Stirrup
Stirrup-401± 20
K2 Horizon 0.9BMBZUAI Institute of Foundation ModelsAgent · Stirrup
Stirrup-402± 20
  1. GDPval-AA v2.1 的 Elo 锚点:DeepSeek V4.1 Flash (Max) 固定为 1600,因此没有置信区间。来源↩

⌘K搜索知识库

最近收录

正在载入知识索引…