Name: 智能体模型厂商排行榜（2026-08-19）- 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	12.34% ±1.53%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 6 确认成功Confirmed Success ▲15.49%±3.15% 赞扬/抱怨Praise vs Complaint ▲19.85%±5.51% 可控性Steerability ▲10.95%±2.98% Bash 恢复Bash Recovery ▲14.31%±0.81% 工具幻觉Tool Hallucination ▼1.09%±0.17% 会话数Sessions 19,783
› 🥈	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	10.53% ±0.62%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 6 确认成功Confirmed Success ▲17.84%±1.24% 赞扬/抱怨Praise vs Complaint ▲18.60%±2.24% 可控性Steerability ▲6.20%±1.13% Bash 恢复Bash Recovery ▲8.79%±0.54% 工具幻觉Tool Hallucination ▼1.19%±0.17% 会话数Sessions 82,558
› 🥉	OpenAIOpenAI GPT 5.6 Sol (xHigh) · Proprietary	9.80% ±1.39%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 11 确认成功Confirmed Success ▲10.12%±2.86% 赞扬/抱怨Praise vs Complaint ▲22.03%±5.09% 可控性Steerability ▲6.01%±2.88% Bash 恢复Bash Recovery ▲9.67%±1.03% 工具幻觉Tool Hallucination ▼1.19%±0.17% 会话数Sessions 26,243
› 4	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	6.29% ±1.39%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 19 确认成功Confirmed Success ▲11.52%±3.10% 赞扬/抱怨Praise vs Complaint ▲6.81%±4.88% 可控性Steerability ▲4.51%±2.73% Bash 恢复Bash Recovery ▲8.41%±1.16% 工具幻觉Tool Hallucination ▼0.20%±0.34% 会话数Sessions 11,483
› 5	SpaceXAISpaceXAI Grok 4.5 · Proprietary	6.08% ±1.15%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 19 确认成功Confirmed Success ▲5.58%±2.57% 赞扬/抱怨Praise vs Complaint ▲5.33%±4.10% 可控性Steerability ▲6.71%±2.17% Bash 恢复Bash Recovery ▲11.61%±1.14% 工具幻觉Tool Hallucination ▼1.19%±0.17% 会话数Sessions 30,806
› 6	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	6.06% ±0.86%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 18 确认成功Confirmed Success ▲7.31%±1.86% 赞扬/抱怨Praise vs Complaint ▲10.14%±3.00% 可控性Steerability ▲5.53%±1.51% Bash 恢复Bash Recovery ▲6.14%±1.12% 工具幻觉Tool Hallucination ▼1.19%±0.17% 会话数Sessions 54,543
› 7	DeepSeekDeepSeek Deepseek V4 Flash (High) (20260731) · MIT	4.15% ±0.79%

综合排名Overall Rank — 名次区间Rank Spread 14 ↔ 23 确认成功Confirmed Success ▲8.61%±1.89% 赞扬/抱怨Praise vs Complaint ▲2.03%±2.66% 可控性Steerability ▲3.75%±1.50% Bash 恢复Bash Recovery ▲5.17%±0.64% 工具幻觉Tool Hallucination ▼1.19%±0.17% 会话数Sessions 41,736
› 8	GoogleGoogle Gemini 3.7 Flash (High) · Proprietary	3.44% ±1.05%

综合排名Overall Rank — 名次区间Rank Spread 16 ↔ 24 确认成功Confirmed Success ▲9.84%±2.50% 赞扬/抱怨Praise vs Complaint ▲0.62%±3.47% 可控性Steerability ▲2.57%±2.08% Bash 恢复Bash Recovery ▲3.00%±1.17% 工具幻觉Tool Hallucination ▼1.17%±0.17% 会话数Sessions 17,407
› 9	MetaMeta Muse Spark 1.1 · Proprietary	0.90% ±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 ▲5.51%±1.81% 可控性Steerability ▲3.46%±1.15% Bash 恢复Bash Recovery ▲5.91%±1.13% 工具幻觉Tool Hallucination ▼1.16%±0.17% 会话数Sessions 75,760
› 10	小米Xiaomi Mimo V2.5 Pro · MIT	2.22% ±0.88%

综合排名Overall Rank — 名次区间Rank Spread 28 ↔ 37 确认成功Confirmed Success ▲3.41%±2.19% 赞扬/抱怨Praise vs Complaint ▲7.77%±2.68% 可控性Steerability ▲1.68%±1.61% Bash 恢复Bash Recovery ▲1.71%±1.45% 工具幻觉Tool Hallucination ▼0.05%±0.32% 会话数Sessions 32,510
› 11	腾讯Tencent Hy3 · Apache 2.0	2.26% ±1.25%

综合排名Overall Rank — 名次区间Rank Spread 27 ↔ 37 确认成功Confirmed Success ▲2.26%±2.87% 赞扬/抱怨Praise vs Complaint ▲1.97%±4.14% 可控性Steerability ▲8.74%±2.37% Bash 恢复Bash Recovery ▲2.99%±1.73% 工具幻觉Tool Hallucination ▼1.29%±0.64% 会话数Sessions 19,581
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	3.14% ±0.83%

综合排名Overall Rank — 名次区间Rank Spread 31 ↔ 37 确认成功Confirmed Success ▲6.61%±2.22% 赞扬/抱怨Praise vs Complaint ▲9.35%±2.53% 可控性Steerability ▲6.31%±1.55% Bash 恢复Bash Recovery ▲5.90%±0.83% 工具幻觉Tool Hallucination ▼0.68%±0.30% 会话数Sessions 32,140
› 13	ThinkyThinky Inkling Small · Apache 2.0	6.69% ±1.76%

综合排名Overall Rank — 名次区间Rank Spread 37 ↔ 42 确认成功Confirmed Success ▲18.02%±4.91% 赞扬/抱怨Praise vs Complaint ▲14.30%±5.38% 可控性Steerability ▲12.41%±3.96% Bash 恢复Bash Recovery ▲10.89%±1.61% 工具幻觉Tool Hallucination ▼0.38%±0.50% 会话数Sessions 8,479
› 14	MistralMistral Mistral Medium 3.5 · Modified MIT	7.52% ±1.89%

综合排名Overall Rank — 名次区间Rank Spread 38 ↔ 44 确认成功Confirmed Success ▲11.26%±4.52% 赞扬/抱怨Praise vs Complaint ▲10.94%±5.41% 可控性Steerability ▲11.58%±3.86% Bash 恢复Bash Recovery ▲1.42%±3.08% 工具幻觉Tool Hallucination ▼2.38%±1.77% 会话数Sessions 5,979
› 15	UpstageUpstage Solar Pro 4 · Proprietary	10.88% ±2.33%

综合排名Overall Rank — 名次区间Rank Spread 40 ↔ 47 确认成功Confirmed Success ▲6.43%±5.68% 赞扬/抱怨Praise vs Complaint ▲17.97%±6.19% 可控性Steerability ▲15.46%±4.76% Bash 恢复Bash Recovery ▲13.51%±4.49% 工具幻觉Tool Hallucination ▼1.03%±3.09% 会话数Sessions 5,172
› 16	NVIDIANvidia Nemotron 3 Ultra · OpenMDW-1.1	15.32% ±2.34%

综合排名Overall Rank — 名次区间Rank Spread 46 ↔ 49 确认成功Confirmed Success ▲15.55%±4.99% 赞扬/抱怨Praise vs Complaint ▲15.24%±6.10% 可控性Steerability ▲20.34%±4.36% Bash 恢复Bash Recovery ▲25.76%±6.85% 工具幻觉Tool Hallucination ▼0.29%±0.42% 会话数Sessions 12,090

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