返回模型榜

7 家独立来源，9 张评测榜单

按问题分配固定预算，同一问题的多张榜单共享该问份额；缺榜只降低确定性，不直接扣分。点击左侧可了解每张榜测什么。

独立来源 7 评测榜单 9 最新数据 2026年9月4日 最近检查 2026年9月4日

Artificial Analysis Intelligence Index

定位: 统一复跑的综合能力指数
数据来源

把多项当前评测放到同一套环境里复跑，得到一个综合能力分。价格只作参考，不进总分。

643 款上榜模型
名次 模型 厂商 原榜分数
#1 claude-fable-5-1 Anthropic 65.7
#2 claude-fable-5-1-xhigh Anthropic 64.8
#3 claude-opus-5 Anthropic 63.1
#4 claude-opus-5-xhigh Anthropic 62.5
#4 claude-fable-5-1-high Anthropic 62.5
#6 muse-spark-1-3 Meta 62.1
#6 claude-fable-5 Anthropic 62.1
#8 claude-opus-5-high Anthropic 61.5
#9 gpt-6-astra OpenAI 61.2
#10 gpt-6-astra-xhigh OpenAI 61.0
#11 grok-4-6 xAI 60.9
#11 gpt-5-6-sol OpenAI 60.9
#13 muse-spark-1-3-xhigh Meta 60.8
#14 claude-fable-5-1-medium Anthropic 60.5
#15 gpt-6-astra-high OpenAI 60.3
#16 grok-4-6-xhigh xAI 60.0
#17 kimi-k3 Moonshot AI 59.7
#18 glm-5-3 Z.ai 59.5
#19 gpt-6-astra-medium OpenAI 59.2
#20 grok-4-6-medium xAI 59.0
#20 gpt-5-6-sol-xhigh OpenAI 59.0
#22 gemini-3-8-flash Google 58.7
#23 claude-opus-5-medium Anthropic 58.6
#24 claude-fable-5-1-low Anthropic 58.1
#24 qwen3-8-max Alibaba 58.1
#26 qwen3-8-2-4t-a95b Alibaba 57.7
#27 glm-5-3-flash Z.ai 57.5
#28 gpt-5-6-sol-high OpenAI 57.3
#28 claude-opus-4-8 Anthropic 57.3
#30 muse-spark-1-2 Meta 56.8
#31 gpt-6-astra-low OpenAI 56.7
#32 gemini-3-8-flash-medium Google 56.6
#32 gpt-5-6-terra OpenAI 56.6
#34 gpt-5-5 OpenAI 56.3
#35 gemini-3-7-flash Google 56.0
#36 grok-4-5 xAI 55.8
#36 qwen3-8-flash-next Alibaba 55.8
#38 gpt-5-6-sol-medium OpenAI 55.6
#39 gpt-6-astra-non-reasoning OpenAI 55.3
#39 claude-sonnet-5 Anthropic 55.3
#41 claude-opus-4-7 Anthropic 55.0
#42 gpt-5-5-high OpenAI 54.7
#43 gemini-3-7-flash-medium Google 53.4
#44 deepseek-v4-pro DeepSeek 53.2
#44 muse-spark-1-1 Meta 53.2
#46 gpt-5-4 OpenAI 53.1
#47 gpt-5-6-terra-xhigh OpenAI 52.8
#48 glm-5-2 Z.ai 52.6
#49 claude-opus-5-low Anthropic 52.5
#50 gpt-5-6-luna OpenAI 52.3
#51 gemini-3-5-flash Google 52.0
#51 qwen3-8-27b Alibaba 52.0
#53 deepseek-v4-flash DeepSeek 51.8
#54 gemini-3-8-flash-low Google 51.7
#54 grok-4-6-low xAI 51.7
#56 gemini-3-6-flash Google 51.6
#57 deepseek-v4-flash-vision DeepSeek 51.5
#58 gpt-5-5-medium OpenAI 51.4
#59 gemini-3-7-flash-low Google 50.9
#60 gpt-5-6-sol-low OpenAI 50.7
#61 gpt-5-6-terra-high OpenAI 50.1
#61 gpt-5-6-luna-xhigh OpenAI 50.1
#63 agnes-2-5-pro-beta Sapiens AI 49.1
#64 claude-sonnet-4-6-adaptive Anthropic 48.4
#65 kimi-k3-low Moonshot AI 48.3
#66 gemini-3-1-pro-preview Google 47.7
#67 motif-3 Motif Technologies 47.4
#68 k2-horizon-375b-a23b MBZUAI Institute of Foundation Models 47.3
#69 gpt-5-6-luna-high OpenAI 47.0
#70 gpt-5-6-terra-medium OpenAI 46.8
#71 gemini-3-5-flash-medium Google 46.7
#71 qwen3-7-max Alibaba 46.7
#73 gpt-5-3-codex OpenAI 45.5
#74 minimax-m3 MiniMax 45.4
#75 deepseek-v4-pro-0424 DeepSeek 45.3
#75 motif-0714 Motif Technologies 45.3
#77 kimi-k2-6 Moonshot AI 45.1
#78 claude-opus-4-6-adaptive Anthropic 44.9
#79 gpt-5-5-low OpenAI 44.5
#79 qwen3-8-27b-medium Alibaba 44.5
#81 muse-spark Meta 44.3
#82 apodex-1-1 Apodex 44.0
#83 claude-opus-4-7-non-reasoning Anthropic 43.9
#84 deepseek-v4-pro-0424-high DeepSeek 43.7
#85 gpt-5-2 OpenAI 43.3
#86 kimi-k2-7-code Moonshot AI 43.0
#86 quasar-438b Multiverse Computing 43.0
#88 qwen3-8-27b-low Alibaba 42.9
#88 mimo-v2-5-pro Xiaomi 42.9
#90 claude-sonnet-5-non-reasoning Anthropic 42.6
#91 inkling 其他 42.3
#92 hy3 Tencent 42.2
#93 deepseek-v4-flash-0420 DeepSeek 42.1
#94 gpt-5-6-sol-non-reasoning OpenAI 41.9
#94 claude-opus-4-5-thinking Anthropic 41.9
#96 nex-n2-pro Nex AGI 41.7
#97 solar-pro4 upstage 41.6
#98 mimo-v2-pro Xiaomi 41.4
#99 gpt-5-6-terra-low OpenAI 41.3
#100 gpt-5-2-codex OpenAI 41.2
#100 inkling-small Thinking Machines 41.2
#102 qwen3-6-max Alibaba 41.1
#103 glm-5-1 Z.ai 41.0
#104 gpt-5-4-mini OpenAI 40.9
#105 grok-build-0-1-06-16 xAI 40.7
#106 gemini-3-pro Google 40.6
#106 glm-5 Z.ai 40.6
#108 qwen3-6-plus Alibaba 40.5
#109 gpt-5-4-low OpenAI 40.2
#110 jt-4-1-flash-236b-a21b China Mobile 39.9
#111 gpt-5-4-nano OpenAI 39.7
#111 agnes-2-5-pro-alpha Sapiens AI 39.7
#113 qwen3-7-plus Alibaba 39.4
#114 glm-5-turbo Z.ai 39.1
#115 deepseek-v4-flash-0420-high DeepSeek 39.0
#116 gpt-5-2-medium OpenAI 38.9
#116 gpt-5-6-luna-medium OpenAI 38.9
#116 minimax-m2-7 MiniMax 38.9
#119 claude-opus-4-6 Anthropic 38.8
#120 gemini-3-flash-reasoning Google 38.7
#121 nvidia-nemotron-3-ultra-550b-a55b NVIDIA 38.3
#122 grok-4-20 xAI 38.0
#122 mimo-v2-5-0424 Xiaomi 38.0
#124 grok-4-3 xAI 37.9
#125 ling-3-0-flash InclusionAI 37.8
#126 qwen3-6-27b Alibaba 37.7
#127 gpt-5-1 OpenAI 37.5
#128 claude-4-5-sonnet-thinking Anthropic 37.4
#128 grok-4-20-0309 xAI 37.4
#128 gemini-3-5-flash-lite Google 37.4
#128 solar-open2-250b Upstage 37.4
#132 mimo-v2-omni-0327 Xiaomi 37.3
#133 gpt-5-codex OpenAI 37.0
#134 grok-4-3-medium xAI 36.9
#135 claude-sonnet-4-6 Anthropic 36.8
#136 grok-4-3-low xAI 36.3
#136 glm-5-1-non-reasoning Z.ai 36.3
#138 kimi-k2-5 Moonshot AI 36.0
#139 mimo-v2-omni Xiaomi 35.9
#140 gemini-3-5-flash-minimal Google 35.8
#140 gpt-5-5-non-reasoning OpenAI 35.8
#142 claude-opus-4-5 Anthropic 35.6
#142 gpt-5-1-codex OpenAI 35.6
#144 kimi-k2-6-non-reasoning Moonshot AI 35.4
#145 gpt-5 OpenAI 35.3
#145 glm-5v-turbo Z.ai 35.3
#147 claude-sonnet-4-6-non-reasoning-low-effort Anthropic 35.1
#147 muse-glimmer Meta 35.1
#149 a-x-k2 SK Telecom 35.0
#150 glm-5-2-non-reasoning Z.ai 34.8
#151 qwen3-8-27b-non-reasoning Alibaba 34.7
#152 gpt-5-medium OpenAI 34.6
#152 gpt-5-6-terra-non-reasoning OpenAI 34.6
#152 qwen3-5-27b Alibaba 34.6
#155 claude-4-1-opus-thinking Anthropic 34.5
#155 glm-4-7 Z.ai 34.5
#155 minimax-m2-5 MiniMax 34.5
#158 hy3-preview Tencent 34.4
#159 qwen3-5-397b-a17b Alibaba 34.3
#159 gpt-5-5-instant-05-26 OpenAI 34.3
#161 grok-4 xAI 34.1
#162 longcat-2-0 Meituan 34.0
#162 g9v3-39a5b AI9Stars 34.0
#162 mimo-v2-0206 Xiaomi 34.0
#165 gemini-3-pro-low Google 33.9
#165 gpt-5-6-luna-low OpenAI 33.9
#167 kat-coder-pro-v2 Kuaishou 33.7
#168 kimi-k2-thinking Moonshot AI 33.5
#169 o3-pro OpenAI 33.3
#170 glm-5-non-reasoning Z.ai 33.2
#171 deepseek-v3-2-reasoning DeepSeek 32.8
#171 qwen3-5-122b-a10b Alibaba 32.8
#173 qwen3-5-397b-a17b-non-reasoning Alibaba 32.7
#174 qwen3-max-thinking Alibaba 32.5
#175 qwen3-6-35b-a3b Alibaba 32.1
#175 minimax-m2-1 MiniMax 32.1
#177 gpt-5-low OpenAI 31.9
#177 deepseek-v4-pro-0424-non-reasoning DeepSeek 31.9
#177 mimo-v2-flash-reasoning Xiaomi 31.9
#180 claude-4-opus-thinking Anthropic 31.7
#180 ring-2-6-1t InclusionAI 31.7
#182 gpt-5-mini-medium OpenAI 31.6
#183 deepseek-v3-1-terminus-reasoning DeepSeek 31.4
#184 qwen3-6-27b-non-reasoning Alibaba 31.3
#184 gpt-5-1-codex-mini OpenAI 31.3
#184 grok-4-1-fast-reasoning xAI 31.3
#184 qwen3-5-omni-plus Alibaba 31.3
#188 o3 OpenAI 31.1
#189 k-exaone-2-0-0803 LG AI Research 31.0
#190 step-3-7-flash StepFun 30.9
#191 gpt-5-4-nano-medium OpenAI 30.8
#192 gpt-5-4-mini-medium OpenAI 30.5
#193 mistral-medium-3-5 Mistral AI 30.4
#194 kimi-k2-5-non-reasoning Moonshot AI 30.1
#195 qwen3-5-27b-non-reasoning Alibaba 30.0
#196 claude-4-5-sonnet Anthropic 29.9
#196 claude-4-5-haiku-reasoning Anthropic 29.9
#196 qwen3-5-35b-a3b Alibaba 29.9
#199 claude-4-sonnet-thinking Anthropic 29.8
#200 gemma-4-31b Google 29.7
#201 deepseek-v4-flash-non-reasoning DeepSeek 29.3
#201 glm-4-6-reasoning Z.ai 29.3
#203 gpt-5-5-instant-06-26 OpenAI 29.2
#204 jt-35b-flash China Mobile 29.0
#205 kat-coder-pro-v1 其他 28.9
#205 minimax-m2 MiniMax 28.9
#207 claude-4-1-opus Anthropic 28.8
#208 mimo-v2-5-pro-non-reasoning Xiaomi 28.4
#209 gpt-5-4-non-reasoning OpenAI 28.3
#210 qwen3-5-122b-a10b-non-reasoning Alibaba 28.2
#211 gemini-3-flash Google 27.9
#211 grok-4-fast-reasoning xAI 27.9
#213 claude-3-7-sonnet-thinking Anthropic 27.6
#214 glm-4-7-non-reasoning Z.ai 27.1
#215 gpt-5-6-luna-non-reasoning OpenAI 26.8
#216 hy3-non-reasoning Tencent 26.6
#216 ling-2-6-1t InclusionAI 26.6
#218 gpt-5-2-non-reasoning OpenAI 26.5
#218 doubao-seed-code ByteDance 26.5
#218 step-3-5-flash StepFun 26.5
#221 o4-mini OpenAI 26.1
#221 gemma-4-26b-a4b Google 26.1
#223 claude-4-opus Anthropic 26.0
#223 claude-4-sonnet Anthropic 26.0
#223 step-3-5-flash-0202 StepFun 26.0
#226 gemini-2-5-pro Google 25.9
#226 deepseek-v3-2-reasoning-0925 DeepSeek 25.9
#228 gpt-5-mini OpenAI 25.8
#229 nvidia-nemotron-3-super-120b-a12b NVIDIA 25.7
#230 gemini-3-1-flash-lite-preview Google 25.6
#231 qwen3-max-thinking-preview Alibaba 25.5
#232 deepseek-v3-2 DeepSeek 25.1
#232 mimo-v2-flash Xiaomi 25.1
#234 grok-4-3-non-reasoning xAI 25.0
#235 qwen3-6-35b-a3b-non-reasoning Alibaba 24.6
#236 ling-3-0-tiny InclusionAI 24.5
#236 qwen3-max Alibaba 24.5
#238 qwen3-5-35b-a3b-non-reasoning Alibaba 24.3
#239 gemini-2-5-flash-preview-09-2025-reasoning Google 24.2
#240 claude-4-5-haiku Anthropic 24.1
#240 gpt-oss-120b OpenAI 24.1
#242 kimi-k2-0905 Moonshot AI 24.0
#243 claude-3-7-sonnet Anthropic 23.9
#243 o1 OpenAI 23.9
#245 granite-4-2-30b IBM 23.7
#246 nemotron-3-5-lightning NVIDIA 23.6
#247 glm-4-6 Z.ai 23.4
#247 gemini-2-5-pro-03-25 Google 23.4
#249 glm-4-7-flash Z.ai 23.3
#250 grok-4-20-0309-non-reasoning xAI 22.9
#250 grok-3-mini-reasoning xAI 22.9
#252 command-a-plus Cohere 22.8
#253 gemini-2-5-pro-05-06 Google 22.7
#254 deepseek-v3-2-speciale DeepSeek 22.6
#255 k-exaone LG AI Research 22.5
#256 gemma-4-31b-non-reasoning Google 22.3
#256 ernie-5-0-thinking-preview Baidu 22.3
#258 grok-4-20-non-reasoning xAI 22.2
#258 gemma-4-12b Google 22.2
#260 nova-2-0-pro-reasoning-medium Amazon 22.1
#261 grok-code-fast-1 xAI 22.0
#262 mercury-2 inception-ai 21.9
#263 qwen3-5-9b Alibaba 21.8
#264 deepseek-v3-1-terminus DeepSeek 21.7
#264 deepseek-v3-2-0925 DeepSeek 21.7
#266 apriel-v1-5-15b-thinker ServiceNow 21.6
#267 deepseek-v3-1 DeepSeek 21.4
#268 nova-2-0-omni-reasoning-medium Amazon 21.3
#268 qwen3-coder-next Alibaba 21.3
#270 deepseek-v3-1-reasoning DeepSeek 21.0
#271 qwen3-vl-235b-a22b-reasoning Alibaba 20.9
#272 nova-2-0-lite-reasoning Amazon 20.8
#272 apriel-v1-6-15b-thinker ServiceNow 20.8
#274 gpt-5-1-non-reasoning OpenAI 20.7
#275 qwen3-5-9b-non-reasoning Alibaba 20.6
#276 exaone-4-5-33b LG AI 20.5
#277 gemma-4-26b-a4b-non-reasoning Google 20.4
#277 deepseek-r1 DeepSeek 20.4
#277 qwen3-5-4b Alibaba 20.4
#280 gemini-2-5-flash-reasoning Google 20.3
#281 north-mini-code Cohere 20.2
#282 gpt-5-nano OpenAI 20.1
#283 qwen3-235b-a22b-instruct-2507-reasoning Alibaba 19.9
#284 nova-2-0-pro-reasoning-low Amazon 19.8
#285 glm-4.5 Z.ai 19.7
#285 mistral-small-4 Mistral AI 19.7
#285 kimi-k2 Moonshot AI 19.7
#288 gpt-4-1 OpenAI 19.6
#288 granite-4-2-8b IBM 19.6
#290 qwen3-max-preview Alibaba 19.4
#291 gpt-5-nano-medium OpenAI 19.2
#291 nova-2-0-lite-reasoning-medium Amazon 19.2
#291 o3-mini OpenAI 19.2
#291 devstral-2 Mistral AI 19.2
#291 qwen3-5-omni-flash Alibaba 19.2
#296 gemini-2-5-flash-preview-09-2025 Google 19.1
#296 o1-pro OpenAI 19.1
#298 jt-mini China Mobile 18.8
#299 trinity-large-thinking 其他 18.7
#300 deepseek-r1-0120 DeepSeek 18.6
#300 grok-3 xAI 18.6
#302 seed-oss-36b-instruct ByteDance 18.5
#303 qwen3-235b-a22b-instruct-2507 Alibaba 18.4
#304 hypernova-60b Multiverse Computing 18.3
#305 qwen3-coder-480b-a35b-instruct Alibaba 18.2
#306 qwen3-vl-32b-reasoning Alibaba 18.1
#307 nova-2-0-lite-reasoning-low Amazon 18.0
#307 magistral-medium-2509 Mistral AI 18.0
#307 sonar-reasoning-pro Perplexity 18.0
#310 minimax-m1-80k MiniMax 17.9
#311 gpt-5-4-nano-non-reasoning OpenAI 17.8
#311 nemotron-cascade-2-30b-a3b NVIDIA 17.8
#313 gemini-2-5-flash-reasoning-04-2025 Google 17.7
#313 devstral-small-2 Mistral AI 17.7
#315 k2-think-v2 LLM360 17.4
#315 longcat-flash-lite Meituan 17.4
#317 gpt-5-minimal OpenAI 17.3
#318 hyperclova-x-seed-think-32b Naver 17.2
#318 o1-preview OpenAI 17.2
#320 grok-4-1-fast xAI 17.0
#321 k-exaone-non-reasoning LG AI Research 16.9
#321 glm-4-6v-reasoning Z.ai 16.9
#321 qwen3-next-80b-a3b-reasoning Alibaba 16.9
#324 gpt-5-4-mini-non-reasoning OpenAI 16.8
#325 nova-2-0-omni-reasoning-low Amazon 16.7
#325 glm-4-5-air Z.ai 16.7
#327 grok-4-fast xAI 16.6
#327 mi-dm-k-2-5-pro-dec28 Korea Telecom 16.6
#329 ring-1t InclusionAI 16.3
#330 g9v3-3b AI9Stars 16.2
#331 qwen3-5-4b-non-reasoning Alibaba 16.1
#332 mistral-large-3 Mistral AI 15.9
#333 o3-mini-high OpenAI 15.7
#333 intellect-3 其他 15.7
#335 glm-4-7-flash-non-reasoning Z.ai 15.6
#336 gpt-5-chatgpt OpenAI 15.4
#337 gpt-oss-20b OpenAI 15.2
#337 gemini-2-5-flash-lite-preview-09-2025-reasoning Google 15.2
#337 solar-open-100b-reasoning Upstage 15.2
#337 deepseek-v3-0324 DeepSeek 15.2
#337 grok-3-reasoning xAI 15.2
#342 nemotron-3-nano-omni-30b-a3b NVIDIA 15.0
#343 gpt-oss-120b-low OpenAI 14.9
#343 mistral-small-3-1 Mistral AI 14.9
#345 gpt-4-1-mini OpenAI 14.8
#346 mistral-medium-3-1 Mistral AI 14.7
#347 qwen3-30b-a3b-2507-reasoning Alibaba 14.6
#348 nvidia-nemotron-3-nano-30b-a3b-reasoning NVIDIA 14.5
#348 llama-4-maverick Meta 14.5
#348 minimax-m1-40k MiniMax 14.5
#348 solar-pro-3 Upstage 14.5
#352 gpt-oss-20b-low OpenAI 14.4
#352 nova-2-0-pro Amazon 14.4
#352 qwen3-vl-235b-a22b-instruct Alibaba 14.4
#355 gpt-5-mini-minimal OpenAI 14.3
#355 granite-4-2-3b IBM 14.3
#357 gemini-2-5-flash Google 14.2
#357 k2-v2 MBZUAI Institute of Foundation Models 14.2
#357 deepseek-v3 DeepSeek 14.2
#357 ling-2-6-flash InclusionAI 14.2
#361 o1-mini OpenAI 14.0
#362 qwen3-next-80b-a3b-instruct Alibaba 13.8
#363 gpt-4-5 OpenAI 13.6
#363 qwen3-coder-30b-a3b-instruct Alibaba 13.6
#363 tri-21b-think-preview Trillion Labs 13.6
#366 qwen3-235b-a22b-instruct-reasoning Alibaba 13.5
#366 diffusiongemma-26b-a4b Google 13.5
#368 qwq-32b Alibaba 13.4
#368 qwen3-vl-30b-a3b-reasoning Alibaba 13.4
#370 gemini-2-0-flash-thinking-exp-0121 Google 13.3
#371 gemma-4-12b-non-reasoning Google 13.2
#372 gemini-2-5-flash-lite-preview-09-2025 Google 13.1
#373 motif-2-12-7b Motif Technologies 12.8
#374 ling-1t InclusionAI 12.7
#374 nova-premier Amazon 12.7
#376 solar-pro-2-preview-reasoning Upstage 12.5
#376 magistral-medium Mistral AI 12.5
#376 mistral-medium-3 Mistral AI 12.5
#379 k2-v2-medium MBZUAI Institute of Foundation Models 12.4
#379 llama-nemotron-super-49b-v1-5-reasoning NVIDIA 12.4
#379 celeris-1 Celeris 12.4
#379 devstral-medium Mistral AI 12.4
#383 mistral-small-4-non-reasoning Mistral AI 12.3
#383 gpt-4o-chatgpt-03-25 OpenAI 12.3
#383 tri-21b-think-v0-5 Trillion Labs 12.3
#386 gemma-4-e4b Google 12.2
#386 llama-3-3-nemotron-super-49b-reasoning NVIDIA 12.2
#386 claude-3-5-haiku Anthropic 12.2
#386 gemini-2-0-flash Google 12.2
#390 sarvam-105b Sarvam AI 11.9
#390 minicpm5-1b OpenBMB 11.9
#390 qwen3-4b-2507-instruct-reasoning Alibaba 11.9
#393 nova-2-0-lite Amazon 11.8
#393 gemini-2-0-pro-experimental-02-05 Google 11.8
#393 claude-3-opus Anthropic 11.8
#393 devstral-small-2505 Mistral AI 11.8
#397 minicpm5-1b-non-reasoning OpenBMB 11.7
#398 gemini-2-5-flash-04-2025 Google 11.6
#398 sonar-reasoning Perplexity 11.6
#400 magistral-small-2509 Mistral AI 11.5
#401 gemini-2-5-flash-lite-reasoning Google 11.4
#401 qwen3-32b-instruct-reasoning Alibaba 11.4
#403 ministral-3-14b Mistral AI 11.2
#404 gpt-4o OpenAI 11.1
#405 deepseek-r1-distill-qwen-32b DeepSeek 11.0
#405 lfm2-5-2-6b Liquid AI 11.0
#405 nanbeige4-1-3b Nanbeige 11.0
#405 qwen3-vl-32b-instruct Alibaba 11.0
#409 glm-4-6v Z.ai 10.9
#410 qwen3-235b-a22b-instruct Alibaba 10.8
#411 mistral-small-3-2 Mistral AI 10.7
#412 gemini-2-0-flash-experimental Google 10.6
#412 magistral-small Mistral AI 10.6
#414 exaone-4-0-32b-reasoning LG AI Research 10.5
#414 qwen3-vl-8b-reasoning Alibaba 10.5
#416 nova-2-0-omni Amazon 10.4
#416 qwen3-14b-instruct-reasoning Alibaba 10.4
#418 llama-4-scout Meta 10.3
#418 deepseek-r1-qwen3-8b DeepSeek 10.3
#420 qwen-2-5-max Alibaba 10.1
#421 solar-pro-2-preview Upstage 9.9
#421 hermes-4-llama-3-1-70b-reasoning Nous Research 9.9
#421 gemini-1-5-pro Google 9.9
#421 qwen3-vl-30b-a3b-instruct Alibaba 9.9
#425 claude-35-sonnet Anthropic 9.8
#425 deepseek-r1-distill-llama-70b DeepSeek 9.8
#427 deepseek-r1-distill-qwen-14b DeepSeek 9.7
#427 falcon-h1r-7b TII 9.7
#429 gpt-4-1-nano OpenAI 9.6
#429 ling-flash-2-0 ant-group 9.6
#431 gemma-4-e2b Google 9.5
#431 qwen3-omni-30b-a3b-reasoning Alibaba 9.5
#433 gpt-4o-2024-08-06 OpenAI 9.4
#433 qwen2-5-72b-instruct Alibaba 9.4
#433 sonar Perplexity 9.4
#436 llama-3-3-instruct-70b Meta 9.3
#436 step-3-vl-10b StepFun 9.3
#438 qwen3-30b-a3b-instruct-reasoning Alibaba 9.2
#439 devstral-small Mistral AI 9.1
#439 QwQ-32B-Preview Alibaba 9.1
#439 sonar-pro Perplexity 9.1
#442 glm-4-5v-reasoning Z.ai 9.0
#442 ministral-3-8b Mistral AI 9.0
#442 mistral-large-2 Mistral AI 9.0
#445 llama-3-1-nemotron-ultra-253b-v1-reasoning NVIDIA 8.9
#445 ernie-4-5-300b-a47b Baidu 8.9
#445 qwen3-30b-a3b-2507 Alibaba 8.9
#448 hermes-4-llama-3-1-405b-reasoning Nous Research 8.8
#448 nvidia-nemotron-nano-12b-v2-vl-reasoning NVIDIA 8.8
#448 solar-pro-2-reasoning Upstage 8.8
#451 gemma-4-e4b-non-reasoning Google 8.7
#451 nvidia-nemotron-nano-9b-v2-reasoning NVIDIA 8.7
#451 granite-4-1-30b IBM 8.7
#454 hermes-4-llama-3-1-405b Nous Research 8.6
#454 gemini-2-0-flash-lite-001 Google 8.6
#454 nvidia-nemotron-3-nano-4b NVIDIA 8.6
#457 llama-nemotron-super-49b-v1-5 NVIDIA 8.5
#457 qwen3-32b-instruct Alibaba 8.5
#459 k2-v2-low MBZUAI Institute of Foundation Models 8.4
#459 llama-3-1-nemotron-nano-4b-reasoning NVIDIA 8.4
#459 gemini-2-0-flash-lite-preview Google 8.4
#459 gpt-4o-2024-05-13 OpenAI 8.4
#459 kimi-linear-48b-a3b-instruct Moonshot AI 8.4
#464 llama-3-3-nemotron-super-49b NVIDIA 8.3
#464 qwen3-8b-instruct-reasoning Alibaba 8.3
#464 llama-3-1-instruct-405b Meta 8.3
#467 qwen3-4b-instruct-reasoning Alibaba 8.2
#467 qwen3-vl-8b-instruct Alibaba 8.2
#469 claude-35-sonnet-june-24 Anthropic 8.1
#469 gpt-4o-chatgpt OpenAI 8.1
#469 lfm2-5-8b-a1b Liquid AI 8.1
#469 tulu3-405b Meta 8.1
#473 ring-flash-2-0 ant-group 8.0
#473 pixtral-large-2411 Mistral AI 8.0
#475 olmo-3-1-32b-think Allen AI 7.9
#476 gpt-5-nano-minimal OpenAI 7.8
#476 gemini-1-5-flash Google 7.8
#476 grok-2-1212 xAI 7.8
#479 qwen3-vl-4b-reasoning Alibaba 7.7
#479 gpt-4-turbo OpenAI 7.7
#481 solar-pro-2 Upstage 7.6
#482 command-a Cohere 7.5
#482 nova-pro Amazon 7.5
#484 qwen3-5-2b Alibaba 7.4
#484 gemma-3-27b Google 7.4
#484 llama-3-1-instruct-8b Meta 7.4
#484 llama-3-1-nemotron-instruct-70b NVIDIA 7.4
#488 grok-beta xAI 7.3
#489 nvidia-nemotron-3-nano-30b-a3b NVIDIA 7.2
#489 nvidia-nemotron-nano-9b-v2 NVIDIA 7.2
#489 qwen2.5-32b-instruct Alibaba 7.2
#492 ministral-3-3b Mistral AI 7.1
#493 mistral-large-2407 Mistral AI 7.0
#494 qwen2-5-coder-32b-instruct Alibaba 6.9
#494 qwen3-4b-2507-instruct Alibaba 6.9
#496 glm-4-5v Z.ai 6.8
#496 qwen3-14b-instruct Alibaba 6.8
#496 gpt-4 OpenAI 6.8
#499 gemini-2-5-flash-lite Google 6.7
#499 hermes-4-llama-3-1-70b Nous Research 6.7
#499 gpt-4o-mini OpenAI 6.7
#499 mistral-small-3 Mistral AI 6.7
#499 nova-lite Amazon 6.7
#504 qwen3-30b-a3b-instruct Alibaba 6.6
#505 qwen3-4b-instruct Alibaba 6.5
#505 deepseek-v2-5 DeepSeek 6.5
#505 llama-3-1-instruct-70b Meta 6.5
#508 sarvam-30b Sarvam AI 6.4
#508 deepseek-v2-5-sep-2024 DeepSeek 6.4
#508 gemini-2-0-flash-thinking-exp-1219 Google 6.4
#508 granite-4-1-8b ibm 6.4
#512 gemma-4-e2b-non-reasoning Google 6.2
#512 deepseek-r1-distill-llama-8b DeepSeek 6.2
#512 mistral-saba Mistral AI 6.2
#512 olmo-3-1-32b-instruct allenai 6.2
#516 gemini-1-5-pro-may-2024 Google 6.1
#516 olmo-3-32b-think Allen AI 6.1
#518 llama-3-2-instruct-90b-vision Meta 6.0
#518 qwen-turbo Alibaba 6.0
#518 r1-1776 Perplexity 6.0
#518 reka-flash Reka AI 6.0
#518 solar-mini Upstage 6.0
#523 grok-1 xAI 5.8
#524 exaone-4-0-32b LG AI Research 5.7
#524 phi-4-mini Microsoft 5.7
#524 qwen2-72b-instruct Alibaba 5.7
#527 gemma-3-12b Google 5.5
#528 qwen3-5-2b-non-reasoning Alibaba 5.3
#529 qwen3-5-0-8b Alibaba 5.2
#529 gemini-1-5-flash-8b Google 5.2
#531 deephermes-3-mistral-24b-preview Nous Research 5.0
#531 jamba-1-7-large AI21 Labs 5.0
#533 granite-4-0-h-small IBM 4.9
#534 qwen3-8b-instruct Alibaba 4.8
#534 hermes-3-llama-3-1-70b Nous Research 4.8
#534 jamba-1-5-large 其他 4.8
#534 qwen3-omni-30b-a3b-instruct Alibaba 4.8
#538 deepseek-coder-v2 DeepSeek 4.7
#538 jamba-1-6-large AI21 Labs 4.7
#538 olmo-2-32b Allen AI 4.7
#541 phi-4 Microsoft 4.6
#541 gemini-1-5-flash-may-2024 Google 4.6
#541 lfm2-24b-a2b Liquid AI 4.6
#544 claude-3-sonnet Anthropic 4.4
#544 granite-4-1-3b IBM 4.4
#544 nova-micro Amazon 4.4
#547 gemini-1-0-ultra Google 4.3
#547 mistral-small Mistral AI 4.3
#547 phi-3-mini Microsoft 4.3
#550 nvidia-nemotron-nano-12b-v2-vl NVIDIA 4.2
#550 gemma-3n-e4b-preview-0520 Google 4.2
#550 phi-4-multimodal Microsoft 4.2
#553 mistral-large Mistral AI 4.1
#553 qwen2-5-coder-7b-instruct Alibaba 4.1
#555 mistral-8x22b-instruct Mistral AI 4.0
#556 llama-2-chat-7b Meta 3.9
#556 llama-3-2-instruct-3b Meta 3.9
#558 jamba-reasoning-3b AI21 Labs 3.8
#558 minicpm-v4-6-1-3b OpenBMB 3.8
#560 qwen1.5-110b-chat Alibaba 3.7
#560 qwen3-vl-4b-instruct Alibaba 3.7
#560 reka-flash-3 Reka 3.7
#563 olmo-3-7b-think Allen AI 3.6
#564 claude-21 Anthropic 3.5
#564 claude-3-haiku Anthropic 3.5
#564 olmo-2-7b Allen AI 3.5
#567 ling-mini-2-0 InclusionAI 3.4
#567 molmo-7b-d Allen Institute for AI 3.4
#569 claude-2 Anthropic 3.3
#569 deepseek-r1-distill-qwen-1-5b DeepSeek 3.3
#569 deepseek-v2 DeepSeek 3.3
#572 gpt-35-turbo OpenAI 3.2
#572 mistral-medium Mistral AI 3.2
#572 mistral-small-2402 Mistral AI 3.2
#575 llama-3-instruct-70b Meta 3.1
#576 arctic-instruct Snowflake 3.0
#576 lfm-40b Liquid AI 3.0
#576 llama-3-2-instruct-11b-vision Meta 3.0
#576 qwen-chat-72b Alibaba 3.0
#580 qwen3-5-0-8b-non-reasoning Alibaba 2.9
#581 palm-2 Google 2.8
#582 deepseek-coder-v2-lite DeepSeek 2.7
#582 gemini-1-0-pro Google 2.7
#584 sarvam-m-reasoning Sarvam 2.6
#584 command-r-plus-04-2024 Cohere 2.6
#584 dbrx 其他 2.6
#584 deepseek-llm-67b-chat DeepSeek 2.6
#584 llama-2-chat-13b Meta 2.6
#584 llama-2-chat-70b Meta 2.6
#584 openchat-35 OpenChat 2.6
#591 exaone-4-0-1-2b-reasoning LG AI Research 2.5
#592 exaone-4-0-1-2b LG AI Research 2.4
#592 olmo-3-7b-instruct Allen AI 2.4
#594 jamba-1-5-mini 其他 2.3
#594 jamba-1-7-mini AI21 Labs 2.3
#594 lfm2-2-6b Liquid AI 2.3
#594 lfm2-5-1-2b-instruct Liquid AI 2.3
#594 lfm2-5-1-2b-thinking Liquid AI 2.3
#599 qwen3-1.7b-instruct-reasoning Alibaba 2.2
#599 granite-4-0-h-nano-1b IBM 2.2
#601 jamba-1-6-mini AI21 Labs 2.1
#602 apertus-70b-instruct Apertus 2.0
#602 gemma-3-270m Google 2.0
#602 granite-4-0-micro IBM 2.0
#602 mixtral-8x7b-instruct Mistral AI 2.0
#606 deephermes-3-llama-3-1-8b-preview Nous Research 1.9
#607 claude-instant Anthropic 1.7
#607 command-r-03-2024 Cohere 1.7
#607 llama-65b Meta 1.7
#607 mistral-7b-instruct Mistral AI 1.7
#607 qwen-chat-14b Alibaba 1.7
#612 granite-4-0-nano-1b IBM 1.6
#612 molmo2-8b allenai 1.6
#614 granite-3-3-8b-instruct IBM 1.3
#614 lfm2-8b-a1b Liquid AI 1.3
#616 qwen3-1.7b-instruct Alibaba 1.1
#617 qwen3-0.6b-instruct Alibaba 1.0
#617 qwen3-0.6b-instruct-reasoning Alibaba 1.0
#617 granite-4-0-350m IBM 1.0
#617 apertus-8b-instruct Apertus 1.0
#617 gemma-3-1b Google 1.0
#617 gemma-3-4b Google 1.0
#617 gemma-3n-e2b Google 1.0
#617 gemma-3n-e4b Google 1.0
#617 granite-4-0-h-350m IBM 1.0
#617 lfm2-1-2b Liquid AI 1.0
#617 lfm2-5-vl-1-6b Liquid AI 1.0
#617 llama-3-2-instruct-1b Meta 1.0
#617 llama-3-instruct-8b Meta 1.0
#617 tiny-aya-global Cohere 1.0
#631 claude-sonnet-5-high Anthropic 0.0
#631 claude-sonnet-5-low Anthropic 0.0
#631 claude-sonnet-5-medium Anthropic 0.0
#631 claude-sonnet-5-xhigh Anthropic 0.0
#631 exaone-4-5-33b-non-reasoning LG AI 0.0
#631 cogito-v2-1-reasoning Deep Cogito 0.0
#631 gpt-5-4-pro OpenAI 0.0
#631 gpt-5-5-pro OpenAI 0.0
#631 gemini-3-deep-think Google 0.0
#631 gpt-3-5-turbo-0613 OpenAI 0.0
#631 gpt-4o-mini-realtime-dec-2024 OpenAI 0.0
#631 gpt-4o-realtime-dec-2024 OpenAI 0.0
#631 midm-250-pro-rsnsft Korea Telecom 0.0

Artificial Analysis Multilingual Index

定位: 中文与多语言能力评测
数据来源

独立评估大模型在中文语义理解、本土知识以及多语言跨语种复杂任务中的综合表现。

该子榜单评测数据滚动更新中，即将收录呈现。

LiveBench Global Average

定位: 滚动客观题库的综合能力总分
数据来源

每月发布防污染新题库，涵盖推理、编程、数学、数据分析、指令遵循与语言任务，严防模型刷榜。

53 款上榜模型
名次 模型 厂商 原榜分数
#1 claude-fable-5-1-max-effort Anthropic 83.4
#2 claude-fable-5-max-effort Anthropic 83.0
#3 muse-spark-1.3-xhigh Meta 81.6
#4 gpt-5.6-sol-max OpenAI 81.1
#5 gpt-5.5-xhigh OpenAI 80.2
#6 claude-opus-5-max-effort Anthropic 80.1
#7 smaug-agentic 其他 79.5
#8 kimi-k3 Moonshot AI 79.2
#9 gemini-3.7-flash-high Google 78.8
#10 qwen3.8-max Alibaba 78.5
#11 grok-4.6 xAI 78.0
#12 gpt-5.4-xhigh OpenAI 78.0
#13 muse-spark-1.2-xhigh Meta 78.0
#14 gpt-5.6-terra-max OpenAI 77.9
#15 deepseek-v4-pro-0813 DeepSeek 77.4
#16 gemini-3.1-pro-preview-high Google 77.0
#17 deepseek-v4-flash-vision-exp DeepSeek 76.8
#18 claude-opus-4-7-xhigh-effort Anthropic 76.5
#19 claude-opus-4-8-max-effort Anthropic 76.2
#20 qwen3.8-flash-next Alibaba 76.2
#21 glm-5.3 Z.ai 76.1
#22 claude-sonnet-5-xhigh-effort Anthropic 76.0
#23 gemini-3.8-flash-high Google 75.8
#24 grok-4.5 xAI 75.8
#25 muse-spark-1.1-xhigh Meta 75.3
#26 qwen3.8-27b Alibaba 75.3
#27 gemini-3.5-flash-high Google 74.6
#28 gpt-5.2-2025-12-11-high OpenAI 74.6
#29 claude-opus-4-6-thinking-auto-high-effort Anthropic 74.5
#30 deepseek-v4-flash-0731 DeepSeek 74.2
#31 gpt-5.2-codex OpenAI 74.0
#32 gemini-3.6-flash-high Google 73.6
#33 gpt-5.6-luna-max OpenAI 73.6
#34 glm-5.2 Z.ai 73.2
#35 qwen3.7-max Alibaba 73.1
#36 claude-sonnet-4-6-thinking-auto-medium-effort Anthropic 73.0
#37 claude-opus-4-5-20251101-thinking-64k-high-effort Anthropic 72.6
#38 inkling-xhigh 其他 71.9
#39 glm-5.3-flash Z.ai 71.6
#40 deepseek-v4-pro DeepSeek 71.6
#41 kimi-k2.6-thinking Moonshot AI 70.5
#42 gpt-5.4-nano-xhigh OpenAI 69.6
#43 ox-alpha-max 其他 69.2
#44 qwen3.6-plus Alibaba 68.9
#45 kimi-k2.7-code Moonshot AI 68.4
#46 grok-build-0.1 xAI 67.8
#47 nemotron-3-ultra-550b-a55b NVIDIA 67.4
#48 minimax-m3 MiniMax 67.3
#49 gpt-5.4-mini-xhigh OpenAI 66.4
#50 deepseek-v4-flash DeepSeek 65.5
#51 qwen3.6-27b Alibaba 64.0
#52 gemini-3.5-flash-lite-high Google 63.9
#53 grok-4.3 xAI 62.2

Arena Text Style-Controlled

定位: 真实用户文本盲选风格受控分
数据来源

通过大规模真实用户盲测对战与 Bradley-Terry 模型计算 Elo，并消除字长和 Markdown 格式偏见的风格受控分。

398 款上榜模型
名次 模型 厂商 原榜分数
#1 claude-fable-5 Anthropic 1507
#2 claude-opus-4-6-high Anthropic 1505
#3 claude-opus-4-7-high Anthropic 1502
#4 muse-spark-1.2 (xHigh) Meta 1499
#5 claude-opus-4-6 Anthropic 1498
#6 claude-opus-4-7 Anthropic 1494
#7 claude-opus-5-high Anthropic 1492
#8 muse-spark-1.1 Meta 1492
#9 gemini-3.7-flash-high Google 1491
#10 kimi-k3-max Moonshot AI 1489
#11 muse-spark Meta 1488
#12 claude-opus-5-max Anthropic 1488
#13 gemini-3.1-pro-preview Google 1487
#14 gemini-3-pro Google 1486
#15 glm-5.3-max Z.ai 1483
#16 gpt-5.6-sol-xhigh OpenAI 1483
#17 gpt-5.5-high OpenAI 1482
#18 claude-opus-4-8-high Anthropic 1482
#19 gemini-3.6-flash-high Google 1480
#20 qwen3.8-max Alibaba 1479
#21 gemini-3.5-flash-high Google 1479
#22 gpt-5.5 OpenAI 1477
#23 gpt-5.4-high OpenAI 1476
#24 gpt-5.2-chat-latest-20260210 OpenAI 1476
#25 grok-4.20-beta1 xAI 1475
#26 gemini-3.5-flash-medium Google 1474
#27 glm-5.3-flash Z.ai 1474
#28 gemini-3-flash Google 1474
#29 qwen3.7-max-preview Alibaba 1474
#30 gpt-5.5-instant OpenAI 1474
#31 claude-opus-4-8 Anthropic 1473
#32 claude-opus-4-5-20251101-high-32k Anthropic 1473
#33 claude-sonnet-4-6 Anthropic 1472
#34 glm-5.2-max Z.ai 1472
#35 grok-4.20-beta-0309-reasoning xAI 1472
#36 grok-4.20-multi-agent-beta-0309 xAI 1471
#37 grok-4.5 xAI 1470
#38 claude-opus-4-5-20251101 Anthropic 1469
#39 mimo-v2.5-pro Xiaomi 1468
#40 ernie-5.1 Baidu 1468
#41 gpt-5.6-terra-xhigh OpenAI 1466
#42 glm-5.1 Z.ai 1466
#43 grok-4.1-thinking xAI 1466
#44 gpt-5.4 OpenAI 1466
#45 qwen3.5-max-preview Alibaba 1466
#46 claude-sonnet-5-high Anthropic 1462
#47 grok-4.6-high xAI 1461
#48 kimi-k2.6 Moonshot AI 1461
#49 qwen3.6-max-preview Alibaba 1460
#50 deepseek-v4-pro-high-20260813 DeepSeek 1460
#51 grok-4.1 xAI 1459
#52 gemini-3-flash (thinking-minimal) Google 1458
#53 glm-5 Z.ai 1458
#54 deepseek-v4-pro DeepSeek 1458
#55 gemini-3.5-flash-lite Google 1457
#56 claude-sonnet-4-5-20250929-high-32k Anthropic 1456
#57 dola-seed-2.0-pro ByteDance 1456
#58 hy3 Tencent 1455
#59 qwen3.7-plus Alibaba 1455
#60 claude-sonnet-4-5-20250929 Anthropic 1455
#61 gpt-5.1-high OpenAI 1455
#62 deepseek-v4-pro-high-preview DeepSeek 1455
#63 gpt-5.6-luna-xhigh OpenAI 1452
#64 gemma-4-31b Google 1451
#65 kimi-k2.5-thinking Moonshot AI 1451
#66 claude-opus-4-1-20250805-thinking-16k Anthropic 1450
#67 gpt-5.3-chat-latest OpenAI 1449
#68 ernie-5.0-preview-1203 Baidu 1449
#69 mimo-v2-pro Xiaomi 1448
#70 gpt-5.4-mini-high OpenAI 1448
#71 claude-opus-4-1-20250805 Anthropic 1448
#72 ernie-5.0-0110 Baidu 1446
#73 gemini-2.5-pro Google 1446
#74 gpt-4.5-preview-2025-02-27 OpenAI 1445
#75 qwen3.6-plus Alibaba 1444
#76 chatgpt-4o-latest-20250326 OpenAI 1443
#77 minimax-m3 MiniMax 1443
#78 grok-4.3 xAI 1443
#79 glm-4.7 Z.ai 1442
#80 qwen3.5-397b-a17b Alibaba 1441
#81 inkling 其他 1439
#82 gpt-5.1 OpenAI 1439
#83 deepseek-v4-flash-high-preview DeepSeek 1438
#84 gemma-4-26b-a4b Google 1438
#85 gpt-5.2-high OpenAI 1438
#86 gpt-5.2 OpenAI 1436
#87 deepseek-v4-flash DeepSeek 1436
#88 longcat-flash-chat-2602-exp meituan 1436
#89 qwen3.8-27b Alibaba 1435
#90 qwen3-max-preview Alibaba 1435
#91 gpt-5-high OpenAI 1434
#92 mimo-v2.5 Xiaomi 1434
#93 glm-5v-turbo Z.ai 1433
#94 gemini-3.1-flash-lite-preview Google 1432
#95 o3-2025-04-16 OpenAI 1431
#96 kimi-k2.5-instant Moonshot AI 1431
#97 grok-4-1-fast-reasoning xAI 1430
#98 kimi-k2-thinking-turbo Moonshot AI 1430
#99 mimo-v2-omni Xiaomi 1430
#100 muse-glimmer Meta 1427
#101 mistral-medium-3.5 Mistral AI 1427
#102 gpt-5-chat OpenAI 1427
#103 nvidia-nemotron-3-ultra-550b-a55b-nvfp4 NVIDIA 1426
#104 claude-opus-4-20250514-thinking-16k Anthropic 1426
#105 amazon-nova-experimental-chat-26-02-10 Amazon 1426
#106 deepseek-v3.2 DeepSeek 1425
#107 deepseek-v3.2-exp-thinking DeepSeek 1425
#108 glm-4.6 Z.ai 1425
#109 qwen3-max-2025-09-23 Alibaba 1424
#110 qwen3-235b-a22b-instruct-2507 Alibaba 1423
#111 deepseek-v3.2-thinking DeepSeek 1423
#112 deepseek-v3.2-exp DeepSeek 1422
#113 deepseek-r1-0528 DeepSeek 1421
#114 ernie-5.0-preview-1022 Baidu 1419
#115 grok-4-fast-chat xAI 1418
#116 kimi-k2-0905-preview Moonshot AI 1418
#117 kimi-k2-0711-preview Moonshot AI 1418
#118 deepseek-v3.1 DeepSeek 1417
#119 qwen3.5-122b-a10b Alibaba 1417
#120 deepseek-v3.1-terminus-thinking DeepSeek 1417
#121 deepseek-v3.1-thinking DeepSeek 1416
#122 minimax-m2.7 MiniMax 1416
#123 amazon-nova-experimental-chat-26-01-10 Amazon 1415
#124 deepseek-v3.1-terminus DeepSeek 1415
#125 gpt-4.1-2025-04-14 OpenAI 1414
#126 claude-opus-4-20250514 Anthropic 1414
#127 qwen3-vl-235b-a22b-instruct Alibaba 1414
#128 mistral-large-3 Mistral AI 1414
#129 claude-haiku-4-5-20251001 Anthropic 1413
#130 hunyuan-hy3-preview Tencent 1413
#131 grok-3-preview-02-24 xAI 1411
#132 glm-4.5 Z.ai 1411
#133 grok-4-0709 xAI 1411
#134 gemini-2.5-flash Google 1410
#135 mistral-medium-2508 Mistral AI 1409
#136 qwen3.5-27b Alibaba 1408
#137 Inkling Small Thinking Machines 1406
#138 grok-4-fast-reasoning xAI 1405
#139 gemini-2.5-flash-preview-09-2025 Google 1404
#140 qwen3-235b-a22b-no-thinking Alibaba 1402
#141 gpt-5.4-nano-high OpenAI 1402
#142 o1-2024-12-17 OpenAI 1402
#143 longcat-flash-chat meituan 1402
#144 claude-sonnet-4-20250514-thinking-32k Anthropic 1401
#145 qwen3-235b-a22b-thinking-2507 Alibaba 1400
#146 qwen3-next-80b-a3b-instruct Alibaba 1399
#147 deepseek-r1 DeepSeek 1398
#148 qwen3.5-flash Alibaba 1397
#149 deepseek-v3-0324 DeepSeek 1396
#150 hunyuan-vision-1.5-thinking Tencent 1395
#151 qwen3.5-35b-a3b Alibaba 1395
#152 qwen3-vl-235b-a22b-thinking Alibaba 1395
#153 amazon-nova-experimental-chat-12-10 Amazon 1395
#154 step-3.5-flash StepFun 1394
#155 mimo-v2-flash (non-thinking) Xiaomi 1392
#156 o4-mini-2025-04-16 OpenAI 1391
#157 minimax-m2.5 MiniMax 1391
#158 claude-sonnet-4-20250514 Anthropic 1390
#159 gpt-5-mini-high OpenAI 1389
#160 o1-preview OpenAI 1388
#161 claude-3-7-sonnet-20250219-thinking-32k Anthropic 1388
#162 hunyuan-t1-20250711 Tencent 1387
#163 mistral-medium-2505 Mistral AI 1387
#164 qwen3-coder-480b-a35b-instruct Alibaba 1387
#165 mimo-v2-flash (thinking) Xiaomi 1386
#166 minimax-m2.1-preview MiniMax 1384
#167 hunyuan-turbos-20250416 Tencent 1383
#168 gpt-4.1-mini-2025-04-14 OpenAI 1382
#169 qwen3-30b-a3b-instruct-2507 Alibaba 1382
#170 gemini-2.5-flash-lite-preview-09-2025-no-thinking Google 1380
#171 trinity-large-preview 其他 1379
#172 glm-4.6v Z.ai 1378
#173 solar-pro4 upstage 1377
#174 qwen3-235b-a22b Alibaba 1375
#175 gemini-2.5-flash-lite-preview-06-17-thinking Google 1375
#176 qwen2.5-max Alibaba 1374
#177 claude-3-5-sonnet-20241022 Anthropic 1374
#178 glm-4.5-air Z.ai 1373
#179 claude-3-7-sonnet-20250219 Anthropic 1372
#180 trinity-large-thinking 其他 1369
#181 qwen3-next-80b-a3b-thinking Alibaba 1369
#182 glm-4.7-flash Z.ai 1367
#183 amazon-nova-experimental-chat-11-10 Amazon 1365
#184 gemma-3-27b-it Google 1365
#185 minimax-m1 MiniMax 1364
#186 grok-3-mini-high xAI 1364
#187 o3-mini-high OpenAI 1363
#188 nvidia-nemotron-3-super-120b-a12b NVIDIA 1361
#189 gemini-2.0-flash-001 Google 1360
#190 deepseek-v3 DeepSeek 1358
#191 grok-3-mini-beta xAI 1358
#192 mistral-small-2506 Mistral AI 1357
#193 intellect-3 其他 1356
#194 command-a-03-2025 Cohere 1354
#195 gemini-2.0-flash-lite-preview-02-05 Google 1353
#196 glm-4.5v Z.ai 1353
#197 gpt-oss-120b OpenAI 1352
#198 nvidia-nemotron-3.5-lightning-30b-a3b-nvfp4 NVIDIA 1352
#199 gemini-1.5-pro-002 Google 1351
#200 amazon-nova-experimental-chat-10-20 Amazon 1350
#201 step-3 StepFun 1350
#202 hunyuan-turbos-20250226 Tencent 1349
#203 o3-mini OpenAI 1348
#204 llama-3.1-nemotron-ultra-253b-v1 NVIDIA 1347
#205 qwen3-32b Alibaba 1347
#206 amazon-nova-experimental-chat-10-09 Amazon 1347
#207 mercury-2 inception-ai 1346
#208 qwen-plus-0125 Alibaba 1346
#209 gpt-4o-2024-05-13 OpenAI 1346
#210 minimax-m2 MiniMax 1346
#211 ling-flash-2.0 ant-group 1344
#212 nvidia-llama-3.3-nemotron-super-49b-v1.5 NVIDIA 1343
#213 claude-3-5-sonnet-20240620 Anthropic 1343
#214 glm-4-plus-0111 Z.ai 1343
#215 gemma-3-12b-it Google 1342
#216 granite-4.2-30b IBM 1341
#217 hunyuan-turbo-0110 Tencent 1341
#218 gpt-5-nano-high OpenAI 1337
#219 o1-mini OpenAI 1337
#220 nova-2-lite Amazon 1336
#221 qwq-32b Alibaba 1336
#222 gemini-advanced-0514 Google 1336
#223 grok-2-2024-08-13 xAI 1335
#224 gpt-4o-2024-08-06 OpenAI 1335
#225 llama-3.1-405b-instruct-bf16 Meta 1335
#226 step-2-16k-exp-202412 StepFun 1333
#227 llama-3.1-405b-instruct-fp8 Meta 1333
#228 olmo-3.1-32b-instruct allenai 1330
#229 yi-lightning 其他 1328
#230 llama-3.3-nemotron-49b-super-v1 NVIDIA 1328
#231 molmo-2-8b allenai 1328
#232 qwen3-30b-a3b Alibaba 1327
#233 llama-4-maverick-17b-128e-instruct Meta 1327
#234 hunyuan-large-2025-02-10 Tencent 1326
#235 gpt-4-turbo-2024-04-09 OpenAI 1324
#236 claude-3-5-haiku-20241022 Anthropic 1324
#237 gemini-1.5-pro-001 Google 1324
#238 deepseek-v2.5-1210 DeepSeek 1323
#239 gpt-4.1-nano-2025-04-14 OpenAI 1322
#240 claude-3-opus-20240229 Anthropic 1322
#241 llama-4-scout-17b-16e-instruct Meta 1321
#242 ring-flash-2.0 ant-group 1321
#243 step-1o-turbo-202506 StepFun 1320
#244 glm-4-plus Z.ai 1319
#245 qwen-max-0919 Alibaba 1318
#246 gpt-4o-mini-2024-07-18 OpenAI 1317
#247 llama-3.3-70b-instruct Meta 1317
#248 gemma-3n-e4b-it Google 1317
#249 gpt-oss-20b OpenAI 1317
#250 qwen2.5-plus-1127 Alibaba 1315
#251 nvidia-nemotron-3-nano-30b-a3b-bf16 NVIDIA 1314
#252 mistral-large-2407 Mistral AI 1314
#253 athene-v2-chat 其他 1314
#254 gpt-4-0125-preview OpenAI 1313
#255 gpt-4-1106-preview OpenAI 1312
#256 hunyuan-standard-2025-02-10 Tencent 1311
#257 gemini-1.5-flash-002 Google 1309
#258 grok-2-mini-2024-08-13 xAI 1308
#259 mercury inception-ai 1308
#260 deepseek-v2.5 DeepSeek 1307
#261 olmo-3-32b-think Allen AI 1307
#262 athene-70b-0725 其他 1306
#263 granite-4.1-8b ibm 1306
#264 mistral-large-2411 Mistral AI 1305
#265 magistral-medium-2506 Mistral AI 1304
#266 gemma-3-4b-it Google 1303
#267 mistral-small-3.1-24b-instruct-2503 Mistral AI 1303
#268 qwen2.5-72b-instruct Alibaba 1303
#269 llama-3.1-nemotron-70b-instruct NVIDIA 1299
#270 granite-4.2-3b IBM 1295
#271 hunyuan-large-vision Tencent 1294
#272 llama-3.1-70b-instruct Meta 1293
#273 amazon-nova-pro-v1.0 Amazon 1290
#274 jamba-1.5-large 其他 1289
#275 gemma-2-27b-it Google 1289
#276 granite-4.2-8b IBM 1288
#277 reka-core-20240904 其他 1288
#278 gpt-4-0314 OpenAI 1287
#279 gemini-1.5-flash-001 Google 1286
#280 llama-3.1-nemotron-51b-instruct NVIDIA 1286
#281 llama-3.1-tulu-3-70b Meta 1286
#282 olmo-3.1-32b-think Allen AI 1286
#283 ibm-granite-h-small ibm 1285
#284 claude-3-sonnet-20240229 Anthropic 1281
#285 gemma-2-9b-it-simpo Google 1280
#286 nemotron-4-340b-instruct NVIDIA 1277
#287 llama-3-70b-instruct Meta 1276
#288 command-r-plus-08-2024 Cohere 1276
#289 gpt-4-0613 OpenAI 1276
#290 mistral-small-24b-instruct-2501 Mistral AI 1274
#291 glm-4-0520 Z.ai 1273
#292 reka-flash-20240904 其他 1272
#293 qwen2.5-coder-32b-instruct Alibaba 1270
#294 c4ai-aya-expanse-32b Cohere 1267
#295 gemma-2-9b-it Google 1267
#296 deepseek-coder-v2 DeepSeek 1265
#297 qwen2-72b-instruct Alibaba 1261
#298 command-r-plus Cohere 1261
#299 claude-3-haiku-20240307 Anthropic 1261
#300 amazon-nova-lite-v1.0 Amazon 1260
#301 gemini-1.5-flash-8b-001 Google 1259
#302 phi-4 Microsoft 1256
#303 olmo-2-0325-32b-instruct allenai 1251
#304 command-r-08-2024 Cohere 1250
#305 mistral-large-2402 Mistral AI 1242
#306 amazon-nova-micro-v1.0 Amazon 1240
#307 jamba-1.5-mini 其他 1239
#308 ministral-8b-2410 Mistral AI 1237
#309 gemini-pro-dev-api Google 1236
#310 qwen1.5-110b-chat Alibaba 1234
#311 hunyuan-standard-256k Tencent 1233
#312 reka-flash-21b-20240226-online 其他 1233
#313 qwen1.5-72b-chat Alibaba 1233
#314 mixtral-8x22b-instruct-v0.1 Mistral AI 1229
#315 command-r Cohere 1226
#316 reka-flash-21b-20240226 其他 1226
#317 gpt-3.5-turbo-0125 OpenAI 1225
#318 llama-3-8b-instruct Meta 1223
#319 gemini-pro Google 1223
#320 c4ai-aya-expanse-8b Cohere 1223
#321 mistral-medium Mistral AI 1222
#322 llama-3.1-tulu-3-8b Meta 1220
#323 yi-1.5-34b-chat 其他 1212
#324 zephyr-orpo-141b-A35b-v0.1 其他 1212
#325 llama-3.1-8b-instruct Meta 1211
#326 granite-3.1-8b-instruct ibm 1208
#327 gpt-3.5-turbo-1106 OpenAI 1203
#328 qwen1.5-32b-chat Alibaba 1203
#329 gemma-2-2b-it Google 1200
#330 phi-3-medium-4k-instruct Microsoft 1197
#331 mixtral-8x7b-instruct-v0.1 Mistral AI 1197
#332 dbrx-instruct-preview 其他 1195
#333 qwen1.5-14b-chat Alibaba 1191
#334 internlm2_5-20b-chat 其他 1190
#335 deepseek-llm-67b-chat DeepSeek 1184
#336 wizardlm-70b Microsoft 1184
#337 yi-34b-chat 其他 1183
#338 granite-3.0-8b-instruct ibm 1183
#339 openchat-3.5 其他 1182
#340 openchat-3.5-0106 其他 1182
#341 gemma-1.1-7b-it Google 1182
#342 snowflake-arctic-instruct 其他 1179
#343 granite-3.1-2b-instruct ibm 1178
#344 tulu-2-dpo-70b 其他 1177
#345 openhermes-2.5-mistral-7b 其他 1175
#346 vicuna-33b 其他 1172
#347 phi-3-small-8k-instruct Microsoft 1171
#348 starling-lm-7b-beta 其他 1171
#349 llama-2-70b-chat Meta 1170
#350 starling-lm-7b-alpha 其他 1167
#351 llama-3.2-3b-instruct Meta 1166
#352 nous-hermes-2-mixtral-8x7b-dpo 其他 1164
#353 granite-3.0-2b-instruct ibm 1156
#354 llama2-70b-steerlm-chat NVIDIA 1154
#355 qwq-32b-preview Alibaba 1154
#356 solar-10.7b-instruct-v1.0 其他 1152
#357 dolphin-2.2.1-mistral-7b 其他 1152
#358 mpt-30b-chat 其他 1150
#359 wizardlm-13b Microsoft 1149
#360 mistral-7b-instruct-v0.2 Mistral AI 1149
#361 falcon-180b-chat 其他 1147
#362 qwen1.5-7b-chat Alibaba 1143
#363 phi-3-mini-4k-instruct-june-2024 Microsoft 1143
#364 llama-2-13b-chat Meta 1141
#365 vicuna-13b 其他 1141
#366 qwen-14b-chat Alibaba 1139
#367 palm-2 Google 1138
#368 gemma-7b-it Google 1137
#369 codellama-34b-instruct Meta 1136
#370 zephyr-7b-beta 其他 1130
#371 phi-3-mini-128k-instruct Microsoft 1129
#372 phi-3-mini-4k-instruct Microsoft 1128
#373 guanaco-33b 其他 1127
#374 zephyr-7b-alpha 其他 1126
#375 stripedhyena-nous-7b 其他 1121
#376 codellama-70b-instruct Meta 1119
#377 gemma-1.1-2b-it Google 1116
#378 vicuna-7b 其他 1115
#379 smollm2-1.7b-instruct 其他 1114
#380 llama-3.2-1b-instruct Meta 1111
#381 mistral-7b-instruct Mistral AI 1110
#382 llama-2-7b-chat Meta 1107
#383 gemma-2b-it Google 1093
#384 qwen1.5-4b-chat Alibaba 1090
#385 olmo-7b-instruct allenai 1073
#386 koala-13b 其他 1070
#387 alpaca-13b 其他 1069
#388 gpt4all-13b-snoozy OpenAI 1067
#389 mpt-7b-chat 其他 1063
#390 chatglm3-6b Z.ai 1056
#391 RWKV-4-Raven-14B 其他 1042
#392 chatglm2-6b Z.ai 1024
#393 oasst-pythia-12b 其他 1023
#394 chatglm-6b 其他 995
#395 fastchat-t5-3b 其他 992
#396 dolly-v2-12b 其他 982
#397 llama-13b Meta 974
#398 stablelm-tuned-alpha-7b Stability AI 953

Arena WebDev

定位: 真实用户网页交付盲选
数据来源

评测模型根据自然语言需求一次性生成交互式前端网页的能力，由真实用户根据视觉呈现与交互效果盲测投票。

121 款上榜模型
名次 模型 厂商 原榜分数
#1 qwen3.8-max-0902 Alibaba 1691
#2 claude-opus-5-max Anthropic 1688
#3 kimi-k3-max Moonshot AI 1674
#4 qwen3.8-max Alibaba 1669
#5 claude-opus-5-high Anthropic 1661
#6 hy4-preview Tencent 1629
#7 grok-4.6-high xAI 1629
#8 claude-fable-5 Anthropic 1628
#9 qwen3.8-flash-next Alibaba 1620
#10 gpt-5.6-sol-xhigh (codex-harness) OpenAI 1616
#11 glm-5.3-max Z.ai 1608
#12 glm-5.3-flash Z.ai 1604
#13 qwen3.8-27b Alibaba 1599
#14 gemini-3.7-flash-high Google 1587
#15 glm-5.2-max Z.ai 1585
#16 deepseek-v4-pro-high-20260813 DeepSeek 1584
#17 deepseek-v4-flash-high DeepSeek 1581
#18 claude-opus-4-8-high Anthropic 1563
#19 claude-opus-4-7 Anthropic 1557
#20 claude-opus-4-7-high Anthropic 1556
#21 grok-4.5 xAI 1555
#22 claude-opus-4-6-high Anthropic 1546
#23 claude-opus-4-8 Anthropic 1540
#24 muse-spark-1.1 Meta 1540
#25 gemini-3.6-flash-high Google 1538
#26 claude-sonnet-5-high Anthropic 1537
#27 claude-opus-4-6 Anthropic 1537
#28 muse-spark-1.2 (xHigh) Meta 1534
#29 seed-2.1-pro-preview ByteDance 1522
#30 claude-sonnet-4-6 Anthropic 1522
#31 gpt-5.6-luna-xhigh (codex-harness) OpenAI 1519
#32 gpt-5.6-terra-xhigh (codex-harness) OpenAI 1518
#33 qwen3.7-max-20260517 Alibaba 1517
#34 hy3 Tencent 1511
#35 gpt-5.5-xhigh (codex-harness) OpenAI 1510
#36 kimi-k2.6 Moonshot AI 1509
#37 glm-5.1 Z.ai 1509
#38 gemini-3.5-flash-high Google 1501
#39 claude-opus-4-5-20251101-high-32k Anthropic 1495
#40 gemini-3.5-flash-medium Google 1492
#41 minimax-m3 MiniMax 1487
#42 gpt-5.5-high (codex-harness) OpenAI 1486
#43 qwen3.6-max-preview Alibaba 1479
#44 mimo-v2.5-pro Xiaomi 1476
#45 kimi-k2.7-code Moonshot AI 1473
#46 claude-opus-4-5-20251101 Anthropic 1468
#47 deepseek-v4-pro-high-preview DeepSeek 1464
#48 gpt-5.4-high (codex-harness) OpenAI 1463
#49 qwen3.6-plus Alibaba 1460
#50 gpt-5.5 (codex-harness) OpenAI 1457
#51 gemini-3.5-flash-lite Google 1449
#52 gemini-3.1-pro-preview Google 1446
#53 deepseek-v4-pro DeepSeek 1445
#54 gpt-5.4-medium (codex-harness) OpenAI 1442
#55 gemini-3-pro Google 1438
#56 mimo-v2.5 Xiaomi 1438
#57 gemini-3-flash Google 1438
#58 kimi-k2.5-thinking Moonshot AI 1436
#59 glm-5 Z.ai 1436
#60 glm-4.7 Z.ai 1435
#61 mimo-v2-pro Xiaomi 1434
#62 deepseek-v4-flash-high-preview DeepSeek 1431
#63 gpt-5-medium OpenAI 1419
#64 gpt-5.2 OpenAI 1416
#65 gpt-5.3-codex (codex-harness) OpenAI 1408
#66 inkling 其他 1408
#67 kimi-k2.5-instant Moonshot AI 1405
#68 Inkling Small Thinking Machines 1403
#69 glm-5v-turbo Z.ai 1400
#70 minimax-m2.7 MiniMax 1399
#71 qwen3.5-397b-a17b Alibaba 1398
#72 gpt-5.4-mini-high OpenAI 1397
#73 claude-sonnet-4-5-20250929-high-32k Anthropic 1393
#74 gpt-5.1-medium OpenAI 1391
#75 claude-opus-4-1-20250805 Anthropic 1389
#76 gpt-5.4 OpenAI 1389
#77 minimax-m2.1-preview MiniMax 1387
#78 claude-sonnet-4-5-20250929 Anthropic 1386
#79 minimax-m2.5 MiniMax 1384
#80 gemini-3-flash (thinking-minimal) Google 1383
#81 grok-4.20-beta-0309-reasoning xAI 1373
#82 solar-pro4 upstage 1370
#84 gemma-4-31b Google 1363
#85 gemma-4-26b-a4b Google 1362
#86 deepseek-v3.2-thinking DeepSeek 1361
#87 muse-glimmer Meta 1359
#88 qwen3.5-122b-a10b Alibaba 1358
#89 qwen3.5-27b Alibaba 1357
#90 hunyuan-hy3-preview Tencent 1356
#91 grok-4.3 xAI 1356
#92 laguna-m.1 poolside 1347
#93 gpt-5.1 OpenAI 1341
#94 glm-4.6 Z.ai 1341
#95 gpt-5.2-codex OpenAI 1339
#96 gpt-5.1-codex OpenAI 1336
#97 mimo-v2-flash (non-thinking) Xiaomi 1331
#98 claude-haiku-4-5-20251001 Anthropic 1329
#99 deepseek-v3.2 DeepSeek 1325
#100 kimi-k2-thinking-turbo Moonshot AI 1323
#101 laguna-xs.2 poolside 1302
#102 minimax-m2 MiniMax 1297
#103 mimo-v2-flash (thinking) Xiaomi 1292
#104 qwen3-coder-480b-a35b-instruct Alibaba 1274
#105 deepseek-v3.2-exp DeepSeek 1272
#106 mistral-medium-3.5 Mistral AI 1265
#107 KAT-Coder-Pro-V1 其他 1254
#108 gemini-3.1-flash-lite-preview Google 1254
#109 qwen3.5-35b-a3b Alibaba 1250
#110 gpt-5.1-codex-mini OpenAI 1243
#111 grok-4-1-fast-reasoning xAI 1241
#112 qwen3.5-flash Alibaba 1238
#113 trinity-large-thinking 其他 1238
#114 mistral-large-3 Mistral AI 1229
#115 gemini-2.5-pro Google 1225
#116 grok-4.1-thinking xAI 1211
#117 devstral-2 Mistral AI 1194
#118 granite-4.1-8b ibm 1192
#119 mercury-2 inception-ai 1166
#120 grok-code-fast-1 xAI 1165
#121 grok-4-fast-reasoning xAI 1162
#122 devstral-medium-2507 Mistral AI 1080

Mercor APEX-Agents

定位: 真实生产环境复杂长任务 Agent 评测
数据来源

由资深专业人士设计、耗时数小时的复杂长链路工作流测试，评测模型作为自主 Agent 的执行完成度与可靠性。

73 款上榜模型
名次 模型 厂商 原榜分数
#1 gpt-6-astra OpenAI 62.4
#2 claude-fable-5.1 Anthropic 62.0
#3 claude-opus-5 Anthropic 60.6
#4 claude-fable-5.1-high Anthropic 60.0
#5 claude-fable-5 Anthropic 59.2
#6 muse-spark-1-1 Meta 58.1
#7 grok-4-6 xAI 57.5
#8 gpt-5-6-sol-max OpenAI 56.7
#9 gpt-5-6-sol-max-pro OpenAI 56.4
#10 claude-opus-4-8 Anthropic 56.2
#11 gpt-5.5-xhigh OpenAI 55.5
#12 kimi-k3 Moonshot AI 55.4
#13 gpt-5-6-sol-xhigh OpenAI 54.3
#14 zai-org/GLM-5.2-FP8 Z.ai 52.2
#15 deepseek-v4-flash DeepSeek 51.6
#16 responses/gpt-5.4 OpenAI 51.3
#17 claude-opus-4-7 Anthropic 50.1
#18 claude-opus-4-6-high Anthropic 48.7
#18 gemini-3.1-pro-preview-high Google 48.7
#20 claude-sonnet-5 Anthropic 48.5
#21 gpt-5.2-xhigh OpenAI 47.8
#22 grok-4-5 xAI 47.1
#23 gemini-3-pro-preview-high Google 46.7
#24 gpt-5.3-codex-high OpenAI 44.6
#25 gpt-5.2-codex-high OpenAI 42.5
#26 kimi-k2-7-code Moonshot AI 41.9
#27 iter-39 Applied Compute 40.1
#28 claude-opus-4-5-high Anthropic 37.1
#29 zai-org/GLM-4.7 Z.ai 21.2
#1 claude-fable-5 Anthropic 60.9
#2 claude-opus-4-8 Anthropic 59.4
#3 gpt-5.5-xhigh OpenAI 53.9
#4 grok-4-5 xAI 50.9
#5 claude-opus-4-7 Anthropic 50.6
#6 claude-opus-4-6-max Anthropic 48.4
#6 gpt-5.2-xhigh OpenAI 48.4
#8 gemini-3.1-pro-preview-high Google 48.2
#9 gpt-5.3-codex-high OpenAI 46.9
#10 claude-opus-4-6-high Anthropic 45.6
#11 gpt-5.2-codex-high OpenAI 42.2
#12 claude-sonnet-4-6-high Anthropic 40.7
#13 gemini-3-flash-preview-high Google 39.5
#14 gpt-5.2-high OpenAI 38.7
#15 responses/galapagos-mini-alpha-xhigh OpenAI 37.5
#16 gpt-5.1-codex-high OpenAI 34.9
#17 claude-opus-4-5-high Anthropic 34.8
#17 gpt-5-codex-high OpenAI 34.8
#19 gemini-3-pro-preview-high Google 34.1
#20 kimi-k2.6 Moonshot AI 33.2
#21 gpt-5-high OpenAI 33.0
#22 gpt-5.1-high OpenAI 31.5
#23 o3-high OpenAI 31.4
#24 zai-org/GLM-5 Z.ai 30.8
#25 grok-4 xAI 30.3
#26 moonshotai/Kimi-K2.5 Moonshot AI 29.2
#27 Qwen/Qwen3.5-397B-A17B-high Alibaba 27.7
#28 nemotron-3-ultra-550b NVIDIA 26.2
#29 responses/galapagos-nano-alpha-xhigh OpenAI 25.5
#30 gemini-3.1-flash-lite-preview-high Google 25.0
#31 grok-4-1-fast-reasoning xAI 24.8
#32 claude-sonnet-4-20250514 Anthropic 23.0
#33 claude-haiku-4-5-high Anthropic 21.4
#34 deepseek-ai/DeepSeek-V3.2 DeepSeek 18.8
#35 MiniMax-M2.5 MiniMax 18.7
#36 gemini-2.5-pro-high Google 17.0
#37 openai/gpt-oss-120b OpenAI 14.5
#38 zai-org/GLM-4.6 Z.ai 11.8
#39 moonshotai/Kimi-K2-Thinking Moonshot AI 11.5
#40 zai-org/GLM-4.7 Z.ai 8.4
#41 grok-3 xAI 7.3
#42 gemini-2.5-flash-high Google 6.4
#43 o1-high OpenAI 5.5
#44 gpt-4o OpenAI 5.4

Vals Finance Agent

定位: 金融分析师实际办公任务自动化评测
数据来源

覆盖财报深度拆解、三表建模、估值计算与投资备忘录撰写等专业金融工作，量化 Agent 在专业办公中的可用性。

56 款上榜模型
名次 模型 厂商 原榜分数
#1 google/gemini-3.8-flash Google 61.4
#2 meta/muse_spark_1_2 Meta 60.6
#3 google/gemini-3.7-flash Google 59.0
#4 anthropic/claude-fable-5-1 Anthropic 58.9
#5 anthropic/claude-opus-5 Anthropic 58.6
#6 google/gemini-3.5-flash Google 57.9
#7 zai/glm-5.3-flash Z.ai 57.9
#8 meta/muse_spark_1_1 Meta 57.2
#9 anthropic/claude-fable-5 Anthropic 56.3
#10 google/gemini-3.6-flash Google 56.3
#11 zai/glm-5.3 Z.ai 55.8
#12 openai/gpt-5.6-luna OpenAI 55.0
#13 openai/gpt-5.6-terra OpenAI 54.4
#14 kimi/kimi-k3 Moonshot AI 54.4
#15 anthropic/claude-opus-4-8 Anthropic 53.9
#16 anthropic/claude-sonnet-5 Anthropic 53.9
#17 openai/gpt-5.6-sol OpenAI 53.8
#18 grok/grok-4.6 xAI 53.7
#19 openai/gpt-6-astra OpenAI 53.5
#20 openai/gpt-5.5 OpenAI 51.8
#21 anthropic/claude-opus-4-7 Anthropic 51.5
#22 anthropic/claude-sonnet-4-6 Anthropic 51.0
#23 alibaba/qwen3.8-max Alibaba 50.6
#24 deepseek/deepseek-v4-pro-0813 DeepSeek 50.4
#25 zai/glm-5.2 Z.ai 49.7
#26 deepseek/deepseek-v4-flash-0731 DeepSeek 49.5
#27 alibaba/qwen3.8-27b Alibaba 48.6
#28 grok/grok-4.5 xAI 48.3
#29 minimax/MiniMax-M3 MiniMax 48.3
#30 alibaba/qwen3.7-max Alibaba 47.8
#31 google/gemini-3.5-flash-lite Google 47.4
#32 thinkingmachines/inkling 其他 46.6
#33 openai/gpt-5.4-mini-2026-03-17 OpenAI 45.4
#34 kimi/kimi-k2.6 Moonshot AI 44.9
#35 zai/glm-5.1 Z.ai 44.8
#36 deepseek/deepseek-v4-pro DeepSeek 44.1
#37 google/gemini-3.1-pro-preview Google 43.0
#38 google/gemini-3-flash-preview Google 42.6
#39 xiaomi/mimo-v2.5-pro Xiaomi 41.5
#40 thinkingmachines/inkling-small Thinking Machines 41.3
#41 alibaba/qwen3.6-plus Alibaba 40.8
#42 alibaba/qwen3.7-plus Alibaba 38.2
#43 openai/gpt-5.4-nano-2026-03-17 OpenAI 38.2
#44 grok/grok-4.3 xAI 37.7
#45 nvidia/nemotron-3-ultra-550b-a55b NVIDIA 37.7
#46 xiaomi/mimo-v2.5 Xiaomi 36.7
#47 kimi/kimi-k2.5-thinking Moonshot AI 35.8
#48 mistralai/mistral-medium-3.5 Mistral AI 32.1
#49 anthropic/claude-haiku-4-5-20251001-thinking Anthropic 31.0
#50 ant/ling-3.0-flash-2607 Ant 30.3
#51 google/gemini-3.1-flash-lite-preview Google 30.0
#52 grok/grok-4.20-0309-reasoning xAI 28.5
#53 minimax/MiniMax-M2.7 MiniMax 27.9
#54 poolside/laguna-m.1 poolside 25.0
#55 fireworks/nemotron-lightning-3p5-30b-a3b NVIDIA 18.5
#56 poolside/laguna-xs.2 poolside 15.6

DeepSWE v1.1

定位: 统一软件工程 Agent 解决率评测
数据来源

在真实工业级开源代码仓库中，给定 Issue 描述，评测模型端到端定位缺陷、编写测试与提交 Patch 的实际解决率。

70 款上榜模型
名次 模型 厂商 原榜分数
#1 gpt-6-astra OpenAI 74.1%
#2 gemini-3-8-flash Google 73.8%
#3 claude-opus-5 Anthropic 73.6%
#4 gpt-6-astra OpenAI 73.2%
#4 gpt-6-astra OpenAI 73.2%
#6 claude-opus-5 Anthropic 73.2%
#7 claude-opus-5 Anthropic 72.8%
#8 gpt-6-astra OpenAI 72.8%
#9 gpt-5-6-sol OpenAI 72.7%
#10 gemini-3-8-flash Google 71.0%
#11 gpt-5-6-sol OpenAI 70.7%
#12 claude-fable-5 Anthropic 69.9%
#13 claude-fable-5 Anthropic 69.7%
#14 gpt-5-6-terra OpenAI 69.6%
#15 gpt-5-6-sol OpenAI 69.4%
#16 glm-5-3 Z.ai 69.0%
#17 claude-opus-5 Anthropic 68.9%
#18 claude-fable-5 Anthropic 68.6%
#19 kimi-k3 Moonshot AI 68.5%
#20 grok-4-6 xAI 67.5%
#21 gpt-5-6-luna OpenAI 67.2%
#22 gpt-6-astra OpenAI 67.0%
#22 gpt-5-5 OpenAI 67.0%
#24 grok-4-6 xAI 66.7%
#25 gemini-3-7-flash Google 65.5%
#26 claude-fable-5 Anthropic 65.4%
#27 gemini-3-7-flash Google 65.3%
#28 grok-4-6 xAI 65.2%
#29 gpt-5-5 OpenAI 64.4%
#30 glm-5-3-flash Z.ai 63.4%
#31 deepseek-v4-pro DeepSeek 62.8%
#32 gpt-5-6-sol OpenAI 61.1%
#33 gpt-5-6-terra OpenAI 60.2%
#34 claude-fable-5 Anthropic 59.6%
#35 claude-opus-4-8 Anthropic 59.0%
#36 claude-opus-5 Anthropic 58.1%
#37 qwen3-8-max Alibaba 57.5%
#38 gpt-5-6-luna OpenAI 56.9%
#39 muse-spark-1-2 Meta 54.9%
#40 claude-opus-4-8 Anthropic 54.4%
#41 gpt-5-5 OpenAI 54.0%
#42 claude-sonnet-5 Anthropic 53.8%
#43 gpt-5-6-terra OpenAI 53.8%
#43 gemini-3-7-flash Google 53.8%
#43 grok-4-5 xAI 53.8%
#46 deepseek-v4-flash DeepSeek 53.3%
#46 muse-spark-1-1 Meta 53.3%
#48 claude-opus-4-8 Anthropic 51.8%
#48 gpt-5-4 OpenAI 51.8%
#50 claude-sonnet-5 Anthropic 49.7%
#51 claude-opus-4-8 Anthropic 48.7%
#52 claude-sonnet-5 Anthropic 48.2%
#53 gemini-3-6-flash Google 46.7%
#54 gpt-5-6-sol OpenAI 45.4%
#55 gpt-5-6-luna OpenAI 44.2%
#56 glm-5-2 Z.ai 43.8%
#57 grok-4-6 xAI 41.6%
#58 claude-opus-4-8 Anthropic 40.8%
#59 claude-sonnet-5 Anthropic 39.8%
#60 glm-5-2 Z.ai 36.3%
#61 gemini-3-5-flash Google 36.1%
#62 gpt-5-6-terra OpenAI 35.1%
#63 kimi-k2-7-code Moonshot AI 30.5%
#64 claude-sonnet-5 Anthropic 30.5%
#65 claude-sonnet-4-6 Anthropic 29.9%
#66 gpt-5-5 OpenAI 27.0%
#67 gpt-5-6-terra OpenAI 24.1%
#68 gemini-3-1-pro-preview Google 11.7%
#69 gpt-5-6-luna OpenAI 11.3%
#70 gpt-5-6-luna OpenAI 1.5%

TapTap Maker Benchmark

定位: 真实游戏引擎 Lua 代码执行与交互评测
数据来源

要求模型根据游戏玩法策划需求生成完整 Lua 逻辑代码，并直接载入游戏引擎运行，根据实体行为与游戏状态严格判定通过率。

29 款上榜模型
名次 模型 厂商 原榜分数
#1 claude-fable-5 Anthropic 89.0%
#2 claude-opus-5 Anthropic 88.9%
#3 grok-4.6 xAI 87.3%
#4 gpt-5.6-sol OpenAI 86.5%
#5 claude-opus-5 Anthropic 86.5%
#6 gemini-3.7-flash Google 84.1%
#7 grok-4.6 xAI 84.1%
#8 glm-5.3 Z.ai 83.6%
#9 claude-opus-4-8 Anthropic 82.8%
#10 qwen3.8-max Alibaba 82.5%
#11 gpt-5.6-terra OpenAI 81.8%
#12 qwen3.8-flash Alibaba 81.5%
#13 glm-5.3-flash Z.ai 81.5%
#14 kimi-k3 Moonshot AI 79.7%
#15 x-ai/grok-4.5 xAI 78.3%
#16 gpt-5.6-luna OpenAI 77.8%
#17 deepseek-v4-pro DeepSeek 77.8%
#18 gemini-3.6-flash Google 76.2%
#19 claude-opus-4-6 Anthropic 76.2%
#20 deepseek-v4-flash DeepSeek 74.6%
#21 glm-5.2 Z.ai 74.3%
#22 claude-sonnet-5 Anthropic 73.3%
#23 hy3 Tencent 73.0%
#24 doubao-seed-evolving ByteDance 72.8%
#25 qwen3.8-27b Alibaba 68.3%
#26 gemini-3.1-pro-preview Google 64.7%
#27 gemini-3.5-flash Google 63.9%
#28 MiniMax/MiniMax-M3 MiniMax 60.1%
#29 claude-haiku-4-5 Anthropic 43.6%