Name: 智能体模型厂商排行榜（2026-09-25）- 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 Fable 5.1 (Max) · Proprietary	12.28% ±3.23%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 6 确认成功Confirmed Success ▲19.69%±4.82% 赞扬/抱怨Praise vs Complaint ▲23.70%±12.98% 可控性Steerability ▲5.54%±6.42% Bash 恢复Bash Recovery ▲12.19%±1.69% 工具幻觉Tool Hallucination ▼0.27%±0.03% 会话数Sessions 6,186
› 🥈	OpenAIOpenAI GPT 6 Astra (Max) · Proprietary	10.85% ±2.29%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 7 确认成功Confirmed Success ▲13.68%±3.53% 赞扬/抱怨Praise vs Complaint ▲34.62%±9.02% 可控性Steerability ▲0.89%±5.11% Bash 恢复Bash Recovery ▲6.46%±1.23% 工具幻觉Tool Hallucination ▼0.36%±0.03% 会话数Sessions 11,353
› 🥉	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	4.55% ±0.54%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 18 确认成功Confirmed Success ▲9.96%±1.10% 赞扬/抱怨Praise vs Complaint ▲8.23%±1.85% 可控性Steerability ▲0.16%±1.13% Bash 恢复Bash Recovery ▲4.34%±0.45% 工具幻觉Tool Hallucination ▼0.36%±0.03% 会话数Sessions 127,610
› 4	腾讯Tencent Hy4 preview · Apache 2.0	4.37% ±0.91%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 20 确认成功Confirmed Success ▲7.71%±1.78% 赞扬/抱怨Praise vs Complaint ▲8.03%±3.26% 可控性Steerability ▲0.45%±1.90% Bash 恢复Bash Recovery ▲6.52%±0.51% 工具幻觉Tool Hallucination ▼0.02%±0.33% 会话数Sessions 50,613
› 5	GoogleGoogle Gemini 3.8 Flash (High) · Proprietary	5.80% ±1.91%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 15 确认成功Confirmed Success ▲14.12%±3.15% 赞扬/抱怨Praise vs Complaint ▲11.32%±7.37% 可控性Steerability ▲1.01%±4.20% Bash 恢复Bash Recovery ▲2.41%±1.57% 工具幻觉Tool Hallucination ▼0.15%±0.15% 会话数Sessions 10,156
› 6	DeepSeekDeepSeek Deepseek V4.1 Flash (Max) · MIT	4.10% ±0.60%

综合排名Overall Rank — 名次区间Rank Spread 9 ↔ 18 确认成功Confirmed Success ▲13.10%±1.14% 赞扬/抱怨Praise vs Complaint ▲3.22%±2.09% 可控性Steerability ▲3.47%±1.42% Bash 恢复Bash Recovery ▲7.42%±0.35% 工具幻觉Tool Hallucination ▼0.22%±0.04% 会话数Sessions 79,760
› 7	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	3.99% ±0.69%

综合排名Overall Rank — 名次区间Rank Spread 9 ↔ 20 确认成功Confirmed Success ▲3.85%±1.46% 赞扬/抱怨Praise vs Complaint ▲10.13%±2.42% 可控性Steerability ▲4.47%±1.32% Bash 恢复Bash Recovery ▲1.14%±0.69% 工具幻觉Tool Hallucination ▼0.36%±0.03% 会话数Sessions 83,150
› 8	MetaMeta Muse Spark 1.3 (Max) · Proprietary	3.90% ±0.60%

综合排名Overall Rank — 名次区间Rank Spread 9 ↔ 20 确认成功Confirmed Success ▲8.49%±1.21% 赞扬/抱怨Praise vs Complaint ▲3.62%±2.07% 可控性Steerability ▲1.67%±1.34% Bash 恢复Bash Recovery ▲5.35%±0.58% 工具幻觉Tool Hallucination ▼0.36%±0.03% 会话数Sessions 59,605
› 9	SpaceXAISpaceXAI Grok 4.7 (xHigh) · Proprietary	3.44% ±4.12%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 32 确认成功Confirmed Success ▲19.96%±5.94% 赞扬/抱怨Praise vs Complaint ▲0.14%±14.63% 可控性Steerability ▲9.40%±10.95% Bash 恢复Bash Recovery ▲6.53%±3.06% 工具幻觉Tool Hallucination ▼0.27%±0.03% 会话数Sessions 2,606
› 10	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	3.36% ±1.28%

综合排名Overall Rank — 名次区间Rank Spread 6 ↔ 25 确认成功Confirmed Success ▲6.69%±2.49% 赞扬/抱怨Praise vs Complaint ▲4.75%±4.67% 可控性Steerability ▲4.17%±2.65% Bash 恢复Bash Recovery ▲2.04%±1.61% 工具幻觉Tool Hallucination ▼0.84%±0.30% 会话数Sessions 19,058
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	6.45% ±0.86%

综合排名Overall Rank — 名次区间Rank Spread 35 ↔ 40 确认成功Confirmed Success ▲10.82%±2.19% 赞扬/抱怨Praise vs Complaint ▲10.40%±2.50% 可控性Steerability ▲4.30%±1.70% Bash 恢复Bash Recovery ▲5.83%±1.52% 工具幻觉Tool Hallucination ▼0.89%±0.26% 会话数Sessions 44,116
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	6.58% ±0.84%

综合排名Overall Rank — 名次区间Rank Spread 35 ↔ 40 确认成功Confirmed Success ▲13.76%±2.27% 赞扬/抱怨Praise vs Complaint ▲12.93%±2.45% 可控性Steerability ▲7.01%±1.71% Bash 恢复Bash Recovery ▲0.87%±0.92% 工具幻觉Tool Hallucination ▼0.08%±0.19% 会话数Sessions 40,758
› 13	ThinkyThinky Inkling Small · Apache 2.0	9.61% ±1.36%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 43 确认成功Confirmed Success ▲21.18%±4.00% 赞扬/抱怨Praise vs Complaint ▲20.23%±3.66% 可控性Steerability ▲11.94%±3.14% Bash 恢复Bash Recovery ▲5.71%±1.23% 工具幻觉Tool Hallucination ▼0.43%±0.33% 会话数Sessions 13,021
› 14	MistralMistral Mistral Medium 3.5 · Modified MIT	11.61% ±1.47%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 43 确认成功Confirmed Success ▲16.07%±3.69% 赞扬/抱怨Praise vs Complaint ▲14.91%±4.05% 可控性Steerability ▲6.85%±3.26% Bash 恢复Bash Recovery ▲17.90%±2.70% 工具幻觉Tool Hallucination ▼2.34%±0.75% 会话数Sessions 10,748
› 15	UpstageUpstage Solar Pro 4 · Proprietary	16.21% ±1.86%

综合排名Overall Rank — 名次区间Rank Spread 44 ↔ 44 确认成功Confirmed Success ▲23.14%±4.89% 赞扬/抱怨Praise vs Complaint ▲20.82%±4.80% 可控性Steerability ▲9.72%±3.67% Bash 恢复Bash Recovery ▲25.45%±4.19% 工具幻觉Tool Hallucination ▼1.93%±0.76% 会话数Sessions 9,245

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