Name: 智能体模型厂商排行榜（2026-08-18）- 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	12.29% ±1.52%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 6 确认成功Confirmed Success ▲15.54%±3.15% 赞扬/抱怨Praise vs Complaint ▲19.50%±5.50% 可控性Steerability ▲10.68%±2.98% Bash 恢复Bash Recovery ▲14.58%±0.81% 工具幻觉Tool Hallucination ▼1.14%±0.18% 会话数Sessions 19,781
› 🥈	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	10.42% ±1.03%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 7 确认成功Confirmed Success ▲15.38%±2.03% 赞扬/抱怨Praise vs Complaint ▲19.73%±3.72% 可控性Steerability ▲7.50%±1.88% Bash 恢复Bash Recovery ▲8.24%±0.85% 工具幻觉Tool Hallucination ▼1.24%±0.18% 会话数Sessions 28,369
› 🥉	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	9.97% ±1.39%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 11 确认成功Confirmed Success ▲10.54%±2.84% 赞扬/抱怨Praise vs Complaint ▲22.07%±5.11% 可控性Steerability ▲6.14%±2.91% Bash 恢复Bash Recovery ▲9.84%±1.04% 工具幻觉Tool Hallucination ▼1.24%±0.18% 会话数Sessions 26,182
› 4	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	6.51% ±1.40%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 18 确认成功Confirmed Success ▲11.24%±3.15% 赞扬/抱怨Praise vs Complaint ▲7.26%±4.93% 可控性Steerability ▲5.04%±2.73% Bash 恢复Bash Recovery ▲8.74%±1.17% 工具幻觉Tool Hallucination ▼0.24%±0.35% 会话数Sessions 11,212
› 5	SpaceXAISpaceXAI Grok 4.5 · Proprietary	6.19% ±1.16%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 18 确认成功Confirmed Success ▲5.58%±2.57% 赞扬/抱怨Praise vs Complaint ▲5.83%±4.12% 可控性Steerability ▲6.40%±2.17% Bash 恢复Bash Recovery ▲11.91%±1.15% 工具幻觉Tool Hallucination ▼1.24%±0.18% 会话数Sessions 30,711
› 6	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	6.09% ±0.85%

综合排名Overall Rank — 名次区间Rank Spread 9 ↔ 18 确认成功Confirmed Success ▲7.42%±1.84% 赞扬/抱怨Praise vs Complaint ▲9.82%±2.96% 可控性Steerability ▲5.63%±1.51% Bash 恢复Bash Recovery ▲6.34%±1.13% 工具幻觉Tool Hallucination ▼1.24%±0.18% 会话数Sessions 54,409
› 7	DeepSeekDeepSeek Deepseek V4 Flash (High) (20260731) · MIT	4.23% ±0.79%

综合排名Overall Rank — 名次区间Rank Spread 16 ↔ 23 确认成功Confirmed Success ▲8.75%±1.90% 赞扬/抱怨Praise vs Complaint ▲1.92%±2.67% 可控性Steerability ▲3.73%±1.51% Bash 恢复Bash Recovery ▲5.53%±0.65% 工具幻觉Tool Hallucination ▼1.24%±0.18% 会话数Sessions 41,045
› 8	GoogleGoogle Gemini 3.7 Flash (High) · Proprietary	3.33% ±1.06%

综合排名Overall Rank — 名次区间Rank Spread 17 ↔ 24 确认成功Confirmed Success ▲9.71%±2.55% 赞扬/抱怨Praise vs Complaint ▲0.22%±3.44% 可控性Steerability ▲2.53%±2.11% Bash 恢复Bash Recovery ▲3.45%±1.19% 工具幻觉Tool Hallucination ▼1.22%±0.18% 会话数Sessions 16,914
› 9	MetaMeta Muse Spark 1.1 · Proprietary	0.99% ±0.61%

综合排名Overall Rank — 名次区间Rank Spread 23 ↔ 28 确认成功Confirmed Success ▲6.42%±1.43% 赞扬/抱怨Praise vs Complaint ▲5.48%±1.81% 可控性Steerability ▲3.54%±1.14% Bash 恢复Bash Recovery ▲6.35%±1.13% 工具幻觉Tool Hallucination ▼1.20%±0.18% 会话数Sessions 75,381
› 10	腾讯Tencent Hy3 · Apache 2.0	1.95% ±1.25%

综合排名Overall Rank — 名次区间Rank Spread 26 ↔ 37 确认成功Confirmed Success ▲1.90%±2.86% 赞扬/抱怨Praise vs Complaint ▲1.37%±4.17% 可控性Steerability ▲8.60%±2.37% Bash 恢复Bash Recovery ▲3.44%±1.72% 工具幻觉Tool Hallucination ▼1.32%±0.66% 会话数Sessions 19,426
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	2.10% ±0.88%

综合排名Overall Rank — 名次区间Rank Spread 28 ↔ 37 确认成功Confirmed Success ▲2.94%±2.17% 赞扬/抱怨Praise vs Complaint ▲7.67%±2.67% 可控性Steerability ▲1.84%±1.61% Bash 恢复Bash Recovery ▲1.85%±1.47% 工具幻觉Tool Hallucination ▼0.08%±0.33% 会话数Sessions 32,382
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	3.08% ±0.83%

综合排名Overall Rank — 名次区间Rank Spread 31 ↔ 37 确认成功Confirmed Success ▲6.35%±2.20% 赞扬/抱怨Praise vs Complaint ▲9.15%±2.52% 可控性Steerability ▲6.70%±1.55% Bash 恢复Bash Recovery ▲6.09%±0.84% 工具幻觉Tool Hallucination ▼0.72%±0.31% 会话数Sessions 32,030
› 13	ThinkyThinky Inkling · Apache 2.0	6.77% ±0.92%

综合排名Overall Rank — 名次区间Rank Spread 38 ↔ 40 确认成功Confirmed Success ▲11.27%±2.62% 赞扬/抱怨Praise vs Complaint ▲17.61%±2.61% 可控性Steerability ▲13.18%±1.93% Bash 恢复Bash Recovery ▲7.55%±1.06% 工具幻觉Tool Hallucination ▼0.65%±0.24% 会话数Sessions 36,466
› 14	MistralMistral Mistral Medium 3.5 · Modified MIT	7.70% ±1.88%

综合排名Overall Rank — 名次区间Rank Spread 38 ↔ 45 确认成功Confirmed Success ▲11.68%±4.55% 赞扬/抱怨Praise vs Complaint ▲11.50%±5.36% 可控性Steerability ▲11.81%±3.86% Bash 恢复Bash Recovery ▲1.08%±3.04% 工具幻觉Tool Hallucination ▼2.41%±1.81% 会话数Sessions 5,941
› 15	UpstageUpstage Solar Pro 4 · Proprietary	10.67% ±2.32%

综合排名Overall Rank — 名次区间Rank Spread 39 ↔ 47 确认成功Confirmed Success ▲6.00%±5.64% 赞扬/抱怨Praise vs Complaint ▲17.84%±6.21% 可控性Steerability ▲15.28%±4.65% Bash 恢复Bash Recovery ▲13.16%±4.52% 工具幻觉Tool Hallucination ▼1.06%±3.19% 会话数Sessions 5,143
› 16	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	15.22% ±2.34%

综合排名Overall Rank — 名次区间Rank Spread 46 ↔ 49 确认成功Confirmed Success ▲15.43%±4.98% 赞扬/抱怨Praise vs Complaint ▲15.05%±6.12% 可控性Steerability ▲20.56%±4.36% Bash 恢复Bash Recovery ▲25.40%±6.82% 工具幻觉Tool Hallucination ▼0.32%±0.43% 会话数Sessions 12,078

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