Name: 智能体模型厂商排行榜（2026-08-27）- 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 Opus 5 (High) · Proprietary	12.99% ±1.71%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 5 确认成功Confirmed Success ▲15.45%±3.56% 赞扬/抱怨Praise vs Complaint ▲19.69%±6.32% 可控性Steerability ▲14.42%±3.25% Bash 恢复Bash Recovery ▲14.45%±0.90% 工具幻觉Tool Hallucination ▼0.94%±0.16% 会话数Sessions 20,468
› 🥈	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	10.24% ±1.48%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 8 确认成功Confirmed Success ▲9.83%±3.07% 赞扬/抱怨Praise vs Complaint ▲21.35%±5.47% 可控性Steerability ▲9.69%±2.92% Bash 恢复Bash Recovery ▲9.28%±1.09% 工具幻觉Tool Hallucination ▼1.02%±0.16% 会话数Sessions 27,506
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	9.22% ±0.61%

综合排名Overall Rank — 名次区间Rank Spread 4 ↔ 8 确认成功Confirmed Success ▲17.88%±1.22% 赞扬/抱怨Praise vs Complaint ▲17.03%±2.15% 可控性Steerability ▲1.99%±1.23% Bash 恢复Bash Recovery ▲8.20%±0.53% 工具幻觉Tool Hallucination ▼1.02%±0.16% 会话数Sessions 92,312
› 4	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	6.66% ±1.18%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 18 确认成功Confirmed Success ▲12.73%±2.49% 赞扬/抱怨Praise vs Complaint ▲8.05%±4.20% 可控性Steerability ▲4.71%±2.40% Bash 恢复Bash Recovery ▲7.77%±1.00% 工具幻觉Tool Hallucination ▼0.07%±0.28% 会话数Sessions 16,510
› 5	DeepSeekDeepSeek DeepSeek V4 Pro (High) (0813) · MIT	6.58% ±1.09%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 18 确认成功Confirmed Success ▲13.63%±2.20% 赞扬/抱怨Praise vs Complaint ▲5.80%±3.69% 可控性Steerability ▲2.50%±2.45% Bash 恢复Bash Recovery ▲9.98%±0.88% 工具幻觉Tool Hallucination ▼1.02%±0.16% 会话数Sessions 20,444
› 6	SpaceXAISpaceXAI Grok 4.5 · Proprietary	6.48% ±1.18%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 18 确认成功Confirmed Success ▲4.99%±2.59% 赞扬/抱怨Praise vs Complaint ▲6.95%±4.18% 可控性Steerability ▲8.68%±2.22% Bash 恢复Bash Recovery ▲10.75%±1.26% 工具幻觉Tool Hallucination ▼1.02%±0.16% 会话数Sessions 32,876
› 7	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	5.94% ±0.79%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 18 确认成功Confirmed Success ▲7.82%±1.74% 赞扬/抱怨Praise vs Complaint ▲9.28%±2.74% 可控性Steerability ▲5.88%±1.49% Bash 恢复Bash Recovery ▲5.72%±0.93% 工具幻觉Tool Hallucination ▼1.02%±0.16% 会话数Sessions 61,300
› 8	GoogleGoogle Gemini 3.7 Flash (High) · Proprietary	1.33% ±0.87%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 31 确认成功Confirmed Success ▲9.82%±2.06% 赞扬/抱怨Praise vs Complaint ▲1.75%±2.81% 可控性Steerability ▲4.42%±1.86% Bash 恢复Bash Recovery ▲2.08%±0.99% 工具幻觉Tool Hallucination ▼0.95%±0.17% 会话数Sessions 24,412
› 9	MetaMeta Muse Spark 1.2 (xHigh) · Proprietary	1.32% ±1.11%

综合排名Overall Rank — 名次区间Rank Spread 21 ↔ 32 确认成功Confirmed Success ▲6.70%±2.74% 赞扬/抱怨Praise vs Complaint ▲6.39%±3.35% 可控性Steerability ▲4.81%±2.39% Bash 恢复Bash Recovery ▲10.09%±1.33% 工具幻觉Tool Hallucination ▼1.01%±0.16% 会话数Sessions 18,407
› 10	腾讯Tencent Hy3 · Apache 2.0	1.19% ±1.28%

综合排名Overall Rank — 名次区间Rank Spread 27 ↔ 42 确认成功Confirmed Success ▲2.86%±2.87% 赞扬/抱怨Praise vs Complaint ▲2.96%±4.07% 可控性Steerability ▲0.46%±2.60% Bash 恢复Bash Recovery ▲0.97%±2.07% 工具幻觉Tool Hallucination ▼1.57%±0.64% 会话数Sessions 22,177
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	2.68% ±0.93%

综合排名Overall Rank — 名次区间Rank Spread 33 ↔ 42 确认成功Confirmed Success ▲3.82%±2.29% 赞扬/抱怨Praise vs Complaint ▲7.44%±2.81% 可控性Steerability ▲2.31%±1.75% Bash 恢复Bash Recovery ▲0.30%±1.47% 工具幻觉Tool Hallucination ▼0.14%±0.34% 会话数Sessions 34,925
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	2.79% ±0.89%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 42 确认成功Confirmed Success ▲5.87%±2.30% 赞扬/抱怨Praise vs Complaint ▲8.82%±2.73% 可控性Steerability ▲5.50%±1.77% Bash 恢复Bash Recovery ▲5.69%±0.83% 工具幻觉Tool Hallucination ▼0.55%±0.26% 会话数Sessions 34,344
› 13	MistralMistral Mistral Medium 3.5 · Modified MIT	5.76% ±1.82%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 46 确认成功Confirmed Success ▲13.60%±4.36% 赞扬/抱怨Praise vs Complaint ▲12.05%±4.98% 可控性Steerability ▲2.10%±4.20% Bash 恢复Bash Recovery ▲2.53%±2.78% 工具幻觉Tool Hallucination ▼2.71%±1.64% 会话数Sessions 6,727
› 14	ThinkyThinky Inkling Small · Apache 2.0	6.18% ±1.89%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 46 确认成功Confirmed Success ▲18.21%±5.24% 赞扬/抱怨Praise vs Complaint ▲18.38%±5.05% 可控性Steerability ▲6.52%±4.96% Bash 恢复Bash Recovery ▲11.94%±1.54% 工具幻觉Tool Hallucination ▼0.27%±0.46% 会话数Sessions 9,191
› 15	UpstageUpstage Solar Pro 4 · Proprietary	9.38% ±3.04%

综合排名Overall Rank — 名次区间Rank Spread 43 ↔ 52 确认成功Confirmed Success ▲7.68%±6.45% 赞扬/抱怨Praise vs Complaint ▲13.96%±8.08% 可控性Steerability ▲4.15%±6.54% Bash 恢复Bash Recovery ▲20.46%±7.63% 工具幻觉Tool Hallucination ▼0.68%±1.91% 会话数Sessions 5,766
› 16	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	12.92% ±2.46%

综合排名Overall Rank — 名次区间Rank Spread 46 ↔ 52 确认成功Confirmed Success ▲15.58%±5.19% 赞扬/抱怨Praise vs Complaint ▲14.62%±6.39% 可控性Steerability ▲7.70%±5.10% Bash 恢复Bash Recovery ▲27.03%±6.89% 工具幻觉Tool Hallucination ▼0.33%±0.34% 会话数Sessions 12,264

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