Name: 智能体模型厂商排行榜（2026-07-28）- 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	12.58% ±2.19%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 7 确认成功Confirmed Success ▲10.26%±4.23% 赞扬/抱怨Praise vs Complaint ▲24.42%±7.82% 可控性Steerability ▲12.93%±4.49% Bash 恢复Bash Recovery ▲13.95%±1.22% 工具幻觉Tool Hallucination ▼1.32%±0.19% 会话数Sessions 23,807
› 🥈	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	10.02% ±1.63%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 11 确认成功Confirmed Success ▲8.61%±3.32% 赞扬/抱怨Praise vs Complaint ▲21.90%±6.10% 可控性Steerability ▲8.96%±3.09% Bash 恢复Bash Recovery ▲9.34%±1.18% 工具幻觉Tool Hallucination ▼1.32%±0.19% 会话数Sessions 16,966
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	9.91% ±1.19%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 10 确认成功Confirmed Success ▲14.23%±2.46% 赞扬/抱怨Praise vs Complaint ▲17.45%±4.26% 可控性Steerability ▲9.67%±2.19% Bash 恢复Bash Recovery ▲6.86%±1.03% 工具幻觉Tool Hallucination ▼1.32%±0.19% 会话数Sessions 19,586
› 4	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	7.08% ±0.96%

综合排名Overall Rank — 名次区间Rank Spread 6 ↔ 15 确认成功Confirmed Success ▲9.08%±1.94% 赞扬/抱怨Praise vs Complaint ▲13.79%±3.42% 可控性Steerability ▲5.96%±1.69% Bash 恢复Bash Recovery ▲5.23%±1.09% 工具幻觉Tool Hallucination ▼1.32%±0.19% 会话数Sessions 43,277
› 5	SpaceXAISpaceXAI Grok 4.5 · Proprietary	5.54% ±1.34%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 20 确认成功Confirmed Success ▲4.65%±2.94% 赞扬/抱怨Praise vs Complaint ▲5.69%±4.70% 可控性Steerability ▲5.14%±2.54% Bash 恢复Bash Recovery ▲10.91%±1.24% 工具幻觉Tool Hallucination ▼1.32%±0.19% 会话数Sessions 24,139
› 6	MetaMeta Muse Spark 1.1 · Proprietary	0.75% ±0.72%

综合排名Overall Rank — 名次区间Rank Spread 20 ↔ 29 确认成功Confirmed Success ▲5.65%±1.67% 赞扬/抱怨Praise vs Complaint ▲4.60%±2.20% 可控性Steerability ▲4.78%±1.38% Bash 恢复Bash Recovery ▲6.17%±1.22% 工具幻觉Tool Hallucination ▼1.30%±0.19% 会话数Sessions 43,216
› 7	GoogleGoogle Gemini 3.1 Pro Preview · Proprietary	0.52% ±0.70%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 31 确认成功Confirmed Success ▲2.69%±1.54% 赞扬/抱怨Praise vs Complaint ▲1.29%±2.21% 可控性Steerability ▲2.37%±1.24% Bash 恢复Bash Recovery ▲10.20%±1.34% 工具幻觉Tool Hallucination ▼1.26%±0.19% 会话数Sessions 72,585
› 8	阿里巴巴Alibaba Qwen3.7 Max · Proprietary	0.53% ±0.97%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 31 确认成功Confirmed Success ▲1.95%±2.38% 赞扬/抱怨Praise vs Complaint ▲5.89%±3.03% 可控性Steerability ▲0.20%±1.76% Bash 恢复Bash Recovery ▲4.61%±1.66% 工具幻觉Tool Hallucination ▼0.80%±0.27% 会话数Sessions 20,914
› 9	DeepSeekDeepSeek DeepSeek V4 Pro · MIT	0.78% ±0.94%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 33 确认成功Confirmed Success ▲3.61%±2.41% 赞扬/抱怨Praise vs Complaint ▲5.37%±3.02% 可控性Steerability ▲0.39%±1.74% Bash 恢复Bash Recovery ▲4.77%±0.98% 工具幻觉Tool Hallucination ▼0.72%±0.26% 会话数Sessions 21,459
› 10	腾讯Tencent Hy3 · Apache 2.0	1.03% ±1.88%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 34 确认成功Confirmed Success ▲1.55%±3.99% 赞扬/抱怨Praise vs Complaint ▲2.83%±6.51% 可控性Steerability ▲8.17%±3.49% Bash 恢复Bash Recovery ▲1.44%±2.58% 工具幻觉Tool Hallucination ▼0.33%±0.52% 会话数Sessions 8,506
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	2.52% ±1.01%

综合排名Overall Rank — 名次区间Rank Spread 26 ↔ 34 确认成功Confirmed Success ▲3.81%±2.46% 赞扬/抱怨Praise vs Complaint ▲8.05%±3.13% 可控性Steerability ▲2.11%±1.89% Bash 恢复Bash Recovery ▲1.09%±1.64% 工具幻觉Tool Hallucination ▼0.30%±0.37% 会话数Sessions 21,399
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	2.55% ±0.93%

综合排名Overall Rank — 名次区间Rank Spread 26 ↔ 34 确认成功Confirmed Success ▲5.67%±2.43% 赞扬/抱怨Praise vs Complaint ▲9.01%±2.95% 可控性Steerability ▲4.56%±1.83% Bash 恢复Bash Recovery ▲5.71%±0.93% 工具幻觉Tool Hallucination ▼0.77%±0.37% 会话数Sessions 20,996
› 13	ThinkyThinky Inkling · Apache 2.0	5.71% ±0.87%

综合排名Overall Rank — 名次区间Rank Spread 35 ↔ 36 确认成功Confirmed Success ▲7.07%±2.35% 赞扬/抱怨Praise vs Complaint ▲15.75%±2.56% 可控性Steerability ▲12.07%±1.80% Bash 恢复Bash Recovery ▲5.92%±1.03% 工具幻觉Tool Hallucination ▼0.41%±0.28% 会话数Sessions 25,699
› 14	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	13.14% ±2.33%

综合排名Overall Rank — 名次区间Rank Spread 40 ↔ 44 确认成功Confirmed Success ▲15.00%±5.18% 赞扬/抱怨Praise vs Complaint ▲13.58%±6.94% 可控性Steerability ▲20.73%±4.93% Bash 恢复Bash Recovery ▲16.58%±5.20% 工具幻觉Tool Hallucination ▼0.19%±0.53% 会话数Sessions 10,820

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