Name: 智能体模型厂商排行榜（2026-07-27）- 最佳 Agent 模型对比 - 厂商聚合排名
Creator: LM Arena
License: https://creativecommons.org/licenses/by/4.0/
Keywords: AI 智能体模型排行榜,智能体 AI 排行榜,最佳 Agent 模型对比,AI 模型排行,智能体评测,净提升

排名Rank ⇕	厂商Lab ⇕
› 🥇	AnthropicAnthropic Claude Fable 5 (High) · Proprietary	12.60% ±2.17%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 5 确认成功Confirmed Success ▲9.93%±4.20% 赞扬/抱怨Praise vs Complaint ▲24.48%±7.77% 可控性Steerability ▲13.05%±4.47% Bash 恢复Bash Recovery ▲14.17%±1.21% 工具幻觉Tool Hallucination ▼1.35%±0.19% 会话数Sessions 23,791
› 🥈	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	10.11% ±1.63%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 9 确认成功Confirmed Success ▲8.49%±3.29% 赞扬/抱怨Praise vs Complaint ▲21.94%±6.12% 可控性Steerability ▲9.20%±3.10% Bash 恢复Bash Recovery ▲9.55%±1.18% 工具幻觉Tool Hallucination ▼1.35%±0.19% 会话数Sessions 16,944
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	9.75% ±1.17%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 9 确认成功Confirmed Success ▲14.83%±2.42% 赞扬/抱怨Praise vs Complaint ▲16.22%±4.17% 可控性Steerability ▲9.17%±2.18% Bash 恢复Bash Recovery ▲7.18%±1.05% 工具幻觉Tool Hallucination ▼1.35%±0.19% 会话数Sessions 19,240
› 4	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	7.12% ±0.95%

综合排名Overall Rank — 名次区间Rank Spread 4 ↔ 14 确认成功Confirmed Success ▲9.07%±1.93% 赞扬/抱怨Praise vs Complaint ▲13.85%±3.43% 可控性Steerability ▲5.86%±1.69% Bash 恢复Bash Recovery ▲5.46%±1.09% 工具幻觉Tool Hallucination ▼1.35%±0.19% 会话数Sessions 42,918
› 5	SpaceXAISpaceXAI Grok 4.5 · Proprietary	5.47% ±1.34%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 18 确认成功Confirmed Success ▲3.82%±2.96% 赞扬/抱怨Praise vs Complaint ▲6.02%±4.72% 可控性Steerability ▲5.11%±2.53% Bash 恢复Bash Recovery ▲11.04%±1.28% 工具幻觉Tool Hallucination ▼1.35%±0.19% 会话数Sessions 23,958
› 6	MetaMeta Muse Spark 1.1 · Proprietary	0.81% ±0.72%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 27 确认成功Confirmed Success ▲5.55%±1.68% 赞扬/抱怨Praise vs Complaint ▲4.44%±2.22% 可控性Steerability ▲4.61%±1.39% Bash 恢复Bash Recovery ▲6.24%±1.27% 工具幻觉Tool Hallucination ▼1.33%±0.19% 会话数Sessions 42,226
› 7	阿里巴巴Alibaba Qwen3.7 Max · Proprietary	0.17% ±0.97%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 29 确认成功Confirmed Success ▲1.17%±2.37% 赞扬/抱怨Praise vs Complaint ▲5.36%±3.08% 可控性Steerability ▲0.34%±1.76% Bash 恢复Bash Recovery ▲5.19%±1.61% 工具幻觉Tool Hallucination ▼0.84%±0.27% 会话数Sessions 20,568
› 8	GoogleGoogle Gemini 3.1 Pro Preview · Proprietary	0.31% ±0.70%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 29 确认成功Confirmed Success ▲2.67%±1.53% 赞扬/抱怨Praise vs Complaint ▲1.52%±2.21% 可控性Steerability ▲2.45%±1.22% Bash 恢复Bash Recovery ▲9.49%±1.28% 工具幻觉Tool Hallucination ▼1.29%±0.20% 会话数Sessions 72,235
› 9	DeepSeekDeepSeek DeepSeek V4 Pro · MIT	0.72% ±0.95%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 31 确认成功Confirmed Success ▲4.01%±2.43% 赞扬/抱怨Praise vs Complaint ▲5.05%±3.06% 可控性Steerability ▲0.45%±1.75% Bash 恢复Bash Recovery ▲5.15%±0.98% 工具幻觉Tool Hallucination ▼0.78%±0.25% 会话数Sessions 21,136
› 10	腾讯Tencent Hy3 · Apache 2.0	1.08% ±1.92%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 32 确认成功Confirmed Success ▲2.28%±4.08% 赞扬/抱怨Praise vs Complaint ▲3.41%±6.70% 可控性Steerability ▲8.16%±3.52% Bash 恢复Bash Recovery ▲1.34%±2.72% 工具幻觉Tool Hallucination ▼0.30%±0.55% 会话数Sessions 8,139
› 11	MiniMaxMiniMax Minimax M3 · MiniMax Community License	2.55% ±0.93%

综合排名Overall Rank — 名次区间Rank Spread 24 ↔ 32 确认成功Confirmed Success ▲6.18%±2.44% 赞扬/抱怨Praise vs Complaint ▲9.17%±2.95% 可控性Steerability ▲4.52%±1.83% Bash 恢复Bash Recovery ▲6.28%±0.86% 工具幻觉Tool Hallucination ▼0.84%±0.37% 会话数Sessions 20,655
› 12	小米Xiaomi Mimo V2.5 Pro · MIT	2.57% ±1.00%

综合排名Overall Rank — 名次区间Rank Spread 24 ↔ 32 确认成功Confirmed Success ▲4.11%±2.47% 赞扬/抱怨Praise vs Complaint ▲8.63%±3.08% 可控性Steerability ▲1.93%±1.90% Bash 恢复Bash Recovery ▲1.35%±1.66% 工具幻觉Tool Hallucination ▼0.49%±0.31% 会话数Sessions 21,068
› 13	ThinkyThinky Inkling · Apache 2.0	5.62% ±0.88%

综合排名Overall Rank — 名次区间Rank Spread 33 ↔ 35 确认成功Confirmed Success ▲6.67%±2.38% 赞扬/抱怨Praise vs Complaint ▲15.95%±2.62% 可控性Steerability ▲12.15%±1.84% Bash 恢复Bash Recovery ▲6.25%±1.02% 工具幻觉Tool Hallucination ▼0.41%±0.29% 会话数Sessions 24,730
› 14	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	13.08% ±2.31%

综合排名Overall Rank — 名次区间Rank Spread 38 ↔ 42 确认成功Confirmed Success ▲14.76%±5.15% 赞扬/抱怨Praise vs Complaint ▲14.01%±6.87% 可控性Steerability ▲20.68%±4.91% Bash 恢复Bash Recovery ▲16.17%±5.19% 工具幻觉Tool Hallucination ▼0.20%±0.53% 会话数Sessions 10,788

没有找到相关模型No matching models found

AI 智能体模型排行榜Agent Models Leaderboard 🏆 综合榜单