Name: 智能体模型厂商排行榜（2026-08-06）- 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 Opus 5 (High) · Proprietary	11.99% ±1.37%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 6 确认成功Confirmed Success ▲16.04%±2.79% 赞扬/抱怨Praise vs Complaint ▲19.46%±5.19% 可控性Steerability ▲10.29%±2.51% Bash 恢复Bash Recovery ▲13.07%±0.78% 工具幻觉Tool Hallucination ▼1.07%±0.16% 会话数Sessions 19,487
› 🥈	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	10.28% ±1.72%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 10 确认成功Confirmed Success ▲9.06%±3.48% 赞扬/抱怨Praise vs Complaint ▲22.70%±6.42% 可控性Steerability ▲8.42%±3.50% Bash 恢复Bash Recovery ▲10.07%±1.23% 工具幻觉Tool Hallucination ▼1.16%±0.16% 会话数Sessions 17,800
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	10.08% ±1.07%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 9 确认成功Confirmed Success ▲14.97%±2.09% 赞扬/抱怨Praise vs Complaint ▲19.68%±3.85% 可控性Steerability ▲7.45%±1.97% Bash 恢复Bash Recovery ▲7.13%±0.89% 工具幻觉Tool Hallucination ▼1.16%±0.16% 会话数Sessions 26,156
› 4	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	6.75% ±0.90%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 17 确认成功Confirmed Success ▲9.21%±1.83% 赞扬/抱怨Praise vs Complaint ▲12.39%±3.21% 可控性Steerability ▲5.62%±1.59% Bash 恢复Bash Recovery ▲5.37%±1.12% 工具幻觉Tool Hallucination ▼1.16%±0.16% 会话数Sessions 50,093
› 5	SpaceXAISpaceXAI Grok 4.5 · Proprietary	5.57% ±1.27%

综合排名Overall Rank — 名次区间Rank Spread 9 ↔ 20 确认成功Confirmed Success ▲5.00%±2.82% 赞扬/抱怨Praise vs Complaint ▲5.29%±4.52% 可控性Steerability ▲4.82%±2.39% Bash 恢复Bash Recovery ▲11.57%±1.09% 工具幻觉Tool Hallucination ▼1.16%±0.16% 会话数Sessions 27,653
› 6	DeepSeekDeepSeek Deepseek V4 Flash (High) (20260731) · MIT	2.84% ±1.10%

综合排名Overall Rank — 名次区间Rank Spread 17 ↔ 22 确认成功Confirmed Success ▲8.53%±2.54% 赞扬/抱怨Praise vs Complaint ▲0.46%±3.64% 可控性Steerability ▲0.43%±2.19% Bash 恢复Bash Recovery ▲3.64%±1.01% 工具幻觉Tool Hallucination ▼1.15%±0.16% 会话数Sessions 18,925
› 7	MetaMeta Muse Spark 1.1 · Proprietary	0.91% ±0.62%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 28 确认成功Confirmed Success ▲6.14%±1.43% 赞扬/抱怨Praise vs Complaint ▲4.86%±1.85% 可控性Steerability ▲3.71%±1.18% Bash 恢复Bash Recovery ▲5.89%±1.28% 工具幻觉Tool Hallucination ▼1.11%±0.16% 会话数Sessions 62,721
› 8	阿里巴巴Alibaba Qwen3.7 Max · Proprietary	0.16% ±0.88%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 31 确认成功Confirmed Success ▲0.24%±2.11% 赞扬/抱怨Praise vs Complaint ▲5.07%±2.72% 可控性Steerability ▲0.21%±1.55% Bash 恢复Bash Recovery ▲4.20%±1.49% 工具幻觉Tool Hallucination ▼0.51%±0.26% 会话数Sessions 27,375
› 9	GoogleGoogle Gemini 3.1 Pro Preview · Proprietary	0.54% ±0.71%

综合排名Overall Rank — 名次区间Rank Spread 23 ↔ 31 确认成功Confirmed Success ▲1.71%±1.57% 赞扬/抱怨Praise vs Complaint ▲2.91%±2.23% 可控性Steerability ▲2.78%±1.23% Bash 恢复Bash Recovery ▲11.13%±1.34% 工具幻觉Tool Hallucination ▼1.01%±0.20% 会话数Sessions 79,360
› 10	腾讯Tencent Hy3 · Apache 2.0	1.12% ±1.42%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 35 确认成功Confirmed Success ▲2.20%±3.04% 赞扬/抱怨Praise vs Complaint ▲3.70%±4.91% 可控性Steerability ▲9.01%±2.60% Bash 恢复Bash Recovery ▲2.82%±1.82% 工具幻觉Tool Hallucination ▼0.89%±0.63% 会话数Sessions 14,807
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	2.42% ±0.92%

综合排名Overall Rank — 名次区间Rank Spread 29 ↔ 36 确认成功Confirmed Success ▲3.88%±2.24% 赞扬/抱怨Praise vs Complaint ▲7.27%±2.82% 可控性Steerability ▲2.19%±1.68% Bash 恢复Bash Recovery ▲1.30%±1.45% 工具幻觉Tool Hallucination ▼0.08%±0.35% 会话数Sessions 27,785
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	2.47% ±0.85%

综合排名Overall Rank — 名次区间Rank Spread 29 ↔ 36 确认成功Confirmed Success ▲5.15%±2.21% 赞扬/抱怨Praise vs Complaint ▲9.10%±2.65% 可控性Steerability ▲4.19%±1.64% Bash 恢复Bash Recovery ▲5.48%±0.82% 工具幻觉Tool Hallucination ▼0.59%±0.33% 会话数Sessions 27,674
› 13	ThinkyThinky Inkling · Apache 2.0	6.60% ±0.89%

综合排名Overall Rank — 名次区间Rank Spread 36 ↔ 39 确认成功Confirmed Success ▲11.06%±2.48% 赞扬/抱怨Praise vs Complaint ▲16.01%±2.60% 可控性Steerability ▲12.57%±1.77% Bash 恢复Bash Recovery ▲6.15%±1.08% 工具幻觉Tool Hallucination ▼0.48%±0.24% 会话数Sessions 32,583
› 14	MistralMistral Mistral Medium 3.5 · Modified MIT	6.82% ±2.16%

综合排名Overall Rank — 名次区间Rank Spread 36 ↔ 42 确认成功Confirmed Success ▲10.43%±5.02% 赞扬/抱怨Praise vs Complaint ▲8.71%±6.47% 可控性Steerability ▲11.31%±4.31% Bash 恢复Bash Recovery ▲0.13%±2.86% 工具幻觉Tool Hallucination ▼3.53%±2.85% 会话数Sessions 4,844
› 15	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	13.75% ±2.33%

综合排名Overall Rank — 名次区间Rank Spread 42 ↔ 46 确认成功Confirmed Success ▲16.20%±5.23% 赞扬/抱怨Praise vs Complaint ▲14.76%±6.61% 可控性Steerability ▲19.08%±4.76% Bash 恢复Bash Recovery ▲19.00%±5.52% 工具幻觉Tool Hallucination ▼0.30%±0.41% 会话数Sessions 11,618

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