Name: 智能体模型厂商排行榜（2026-08-25）- 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	12.73% ±1.63%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 5 确认成功Confirmed Success ▲15.97%±3.34% 赞扬/抱怨Praise vs Complaint ▲19.13%±5.89% 可控性Steerability ▲13.36%±3.16% Bash 恢复Bash Recovery ▲14.22%±0.84% 工具幻觉Tool Hallucination ▼0.99%±0.17% 会话数Sessions 20,070
› 🥈	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	10.31% ±1.45%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 8 确认成功Confirmed Success ▲10.02%±2.99% 赞扬/抱怨Praise vs Complaint ▲22.22%±5.29% 可控性Steerability ▲9.16%±2.93% Bash 恢复Bash Recovery ▲9.06%±1.14% 工具幻觉Tool Hallucination ▼1.08%±0.17% 会话数Sessions 26,960
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	9.53% ±0.61%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 8 确认成功Confirmed Success ▲18.24%±1.20% 赞扬/抱怨Praise vs Complaint ▲17.56%±2.15% 可控性Steerability ▲2.31%±1.24% Bash 恢复Bash Recovery ▲8.47%±0.52% 工具幻觉Tool Hallucination ▼1.08%±0.17% 会话数Sessions 89,758
› 4	DeepSeekDeepSeek DeepSeek V4 Pro (High) (0813) · MIT	6.97% ±1.11%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 17 确认成功Confirmed Success ▲13.49%±2.26% 赞扬/抱怨Praise vs Complaint ▲6.38%±3.79% 可控性Steerability ▲3.16%±2.46% Bash 恢复Bash Recovery ▲10.74%±0.76% 工具幻觉Tool Hallucination ▼1.08%±0.17% 会话数Sessions 18,844
› 5	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	6.46% ±1.23%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 19 确认成功Confirmed Success ▲12.52%±2.65% 赞扬/抱怨Praise vs Complaint ▲6.81%±4.36% 可控性Steerability ▲4.66%±2.51% Bash 恢复Bash Recovery ▲8.22%±1.03% 工具幻觉Tool Hallucination ▼0.08%±0.30% 会话数Sessions 14,712
› 6	SpaceXAISpaceXAI Grok 4.5 · Proprietary	6.38% ±1.18%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 19 确认成功Confirmed Success ▲4.72%±2.60% 赞扬/抱怨Praise vs Complaint ▲6.65%±4.18% 可控性Steerability ▲8.49%±2.22% Bash 恢复Bash Recovery ▲10.98%±1.30% 工具幻觉Tool Hallucination ▼1.08%±0.17% 会话数Sessions 32,149
› 7	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	5.86% ±0.83%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 20 确认成功Confirmed Success ▲7.25%±1.80% 赞扬/抱怨Praise vs Complaint ▲8.97%±2.83% 可控性Steerability ▲6.12%±1.56% Bash 恢复Bash Recovery ▲5.89%±1.01% 工具幻觉Tool Hallucination ▼1.08%±0.17% 会话数Sessions 58,668
› 8	MetaMeta Muse Spark 1.2 (xHigh) · Proprietary	1.60% ±1.10%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 31 确认成功Confirmed Success ▲7.06%±2.70% 赞扬/抱怨Praise vs Complaint ▲6.19%±3.35% 可控性Steerability ▲4.34%±2.36% Bash 恢复Bash Recovery ▲10.42%±1.29% 工具幻觉Tool Hallucination ▼1.07%±0.17% 会话数Sessions 18,406
› 9	GoogleGoogle Gemini 3.7 Flash (High) · Proprietary	1.57% ±0.88%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 30 确认成功Confirmed Success ▲9.97%±2.07% 赞扬/抱怨Praise vs Complaint ▲2.04%±2.78% 可控性Steerability ▲3.59%±1.91% Bash 恢复Bash Recovery ▲2.47%±0.98% 工具幻觉Tool Hallucination ▼1.01%±0.17% 会话数Sessions 24,091
› 10	腾讯Tencent Hy3 · Apache 2.0	0.42% ±1.30%

综合排名Overall Rank — 名次区间Rank Spread 24 ↔ 36 确认成功Confirmed Success ▲1.68%±2.82% 赞扬/抱怨Praise vs Complaint ▲0.79%±4.24% 可控性Steerability ▲0.78%±2.64% Bash 恢复Bash Recovery ▲1.02%±2.13% 工具幻觉Tool Hallucination ▼1.44%±0.61% 会话数Sessions 21,290
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	2.79% ±0.91%

综合排名Overall Rank — 名次区间Rank Spread 33 ↔ 42 确认成功Confirmed Success ▲3.50%±2.25% 赞扬/抱怨Praise vs Complaint ▲8.15%±2.71% 可控性Steerability ▲2.67%±1.74% Bash 恢复Bash Recovery ▲0.41%±1.49% 工具幻觉Tool Hallucination ▼0.02%±0.32% 会话数Sessions 34,082
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	3.10% ±0.87%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 42 确认成功Confirmed Success ▲6.61%±2.29% 赞扬/抱怨Praise vs Complaint ▲9.60%±2.60% 可控性Steerability ▲5.61%±1.75% Bash 恢复Bash Recovery ▲5.69%±0.84% 工具幻觉Tool Hallucination ▼0.63%±0.27% 会话数Sessions 33,558
› 13	MistralMistral Mistral Medium 3.5 · Modified MIT	5.22% ±1.85%

综合排名Overall Rank — 名次区间Rank Spread 36 ↔ 44 确认成功Confirmed Success ▲11.91%±4.41% 赞扬/抱怨Praise vs Complaint ▲12.18%±5.09% 可控性Steerability ▲2.73%±4.19% Bash 恢复Bash Recovery ▲2.11%±2.92% 工具幻觉Tool Hallucination ▼2.64%±1.80% 会话数Sessions 6,351
› 14	ThinkyThinky Inkling Small · Apache 2.0	5.42% ±1.85%

综合排名Overall Rank — 名次区间Rank Spread 37 ↔ 44 确认成功Confirmed Success ▲17.40%±5.09% 赞扬/抱怨Praise vs Complaint ▲16.12%±5.27% 可控性Steerability ▲5.25%±4.74% Bash 恢复Bash Recovery ▲11.49%±1.64% 工具幻觉Tool Hallucination ▼0.19%±0.56% 会话数Sessions 8,827
› 15	UpstageUpstage Solar Pro 4 · Proprietary	8.79% ±2.64%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 50 确认成功Confirmed Success ▲6.24%±6.11% 赞扬/抱怨Praise vs Complaint ▲15.29%±7.42% 可控性Steerability ▲8.52%±6.22% Bash 恢复Bash Recovery ▲12.92%±4.38% 工具幻觉Tool Hallucination ▼0.95%±2.33% 会话数Sessions 5,437
› 16	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	12.67% ±2.43%

综合排名Overall Rank — 名次区间Rank Spread 44 ↔ 50 确认成功Confirmed Success ▲15.58%±5.09% 赞扬/抱怨Praise vs Complaint ▲14.23%±6.30% 可控性Steerability ▲7.56%±5.06% Bash 恢复Bash Recovery ▲26.33%±6.80% 工具幻觉Tool Hallucination ▼0.35%±0.35% 会话数Sessions 12,257

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