Name: 智能体模型厂商排行榜（2026-09-28）- 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	14.06% ±1.77%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 3 确认成功Confirmed Success ▲17.37%±2.77% 赞扬/抱怨Praise vs Complaint ▲32.43%±7.15% 可控性Steerability ▲8.40%±3.74% Bash 恢复Bash Recovery ▲11.78%±0.84% 工具幻觉Tool Hallucination ▼0.34%±0.03% 会话数Sessions 15,125
› 🥈	OpenAIOpenAI GPT 6 Astra (Max) · Proprietary	10.36% ±2.28%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 8 确认成功Confirmed Success ▲12.68%±3.58% 赞扬/抱怨Praise vs Complaint ▲33.92%±8.96% 可控性Steerability ▲1.39%±5.07% Bash 恢复Bash Recovery ▲6.26%±1.25% 工具幻觉Tool Hallucination ▼0.35%±0.03% 会话数Sessions 11,547
› 🥉	腾讯Tencent Hy4 preview · Apache 2.0	4.10% ±0.87%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 21 确认成功Confirmed Success ▲7.53%±1.69% 赞扬/抱怨Praise vs Complaint ▲7.45%±3.13% 可控性Steerability ▲0.70%±1.82% Bash 恢复Bash Recovery ▲6.22%±0.51% 工具幻觉Tool Hallucination ▼0.02%±0.32% 会话数Sessions 54,471
› 4	月之暗面Moonshot Kimi K3 (Max) · Kimi K3 license	4.03% ±0.53%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 18 确认成功Confirmed Success ▲9.18%±1.10% 赞扬/抱怨Praise vs Complaint ▲7.35%±1.80% 可控性Steerability ▲0.54%±1.12% Bash 恢复Bash Recovery ▲3.83%±0.50% 工具幻觉Tool Hallucination ▼0.35%±0.03% 会话数Sessions 130,636
› 5	SpaceXAISpaceXAI Grok 4.7 (xHigh) · Proprietary	4.02% ±1.57%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 23 确认成功Confirmed Success ▲8.99%±3.06% 赞扬/抱怨Praise vs Complaint ▲4.42%±5.37% 可控性Steerability ▲0.46%±3.68% Bash 恢复Bash Recovery ▲5.88%±1.03% 工具幻觉Tool Hallucination ▼0.35%±0.03% 会话数Sessions 11,004
› 6	DeepSeekDeepSeek Deepseek V4.1 Flash (Max) · MIT	3.86% ±0.57%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 21 确认成功Confirmed Success ▲12.08%±1.11% 赞扬/抱怨Praise vs Complaint ▲2.90%±1.97% 可控性Steerability ▲3.29%±1.34% Bash 恢复Bash Recovery ▲7.39%±0.33% 工具幻觉Tool Hallucination ▼0.22%±0.04% 会话数Sessions 90,275
› 7	智谱 ZAIZ.ai GLM 5.2 (Max) · MIT · SiliconFlow	3.84% ±0.69%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 21 确认成功Confirmed Success ▲3.78%±1.45% 赞扬/抱怨Praise vs Complaint ▲9.84%±2.40% 可控性Steerability ▲4.18%±1.33% Bash 恢复Bash Recovery ▲1.08%±0.69% 工具幻觉Tool Hallucination ▼0.35%±0.03% 会话数Sessions 84,114
› 8	MetaMeta Muse Spark 1.3 (Max) · Proprietary	3.83% ±0.59%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 21 确认成功Confirmed Success ▲8.35%±1.19% 赞扬/抱怨Praise vs Complaint ▲3.67%±2.00% 可控性Steerability ▲1.58%±1.29% Bash 恢复Bash Recovery ▲5.20%±0.57% 工具幻觉Tool Hallucination ▼0.35%±0.03% 会话数Sessions 63,901
› 9	GoogleGoogle Gemini 3.8 Flash (High) · Proprietary	5.38% ±1.75%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 19 确认成功Confirmed Success ▲14.50%±2.94% 赞扬/抱怨Praise vs Complaint ▲8.18%±6.72% 可控性Steerability ▲2.25%±3.86% Bash 恢复Bash Recovery ▲1.81%±1.47% 工具幻觉Tool Hallucination ▼0.17%±0.13% 会话数Sessions 11,262
› 10	阿里巴巴Alibaba Qwen3.8 Max · Proprietary	3.08% ±1.25%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 27 确认成功Confirmed Success ▲6.36%±2.42% 赞扬/抱怨Praise vs Complaint ▲4.64%±4.55% 可控性Steerability ▲3.38%±2.61% Bash 恢复Bash Recovery ▲1.85%±1.58% 工具幻觉Tool Hallucination ▼0.82%±0.29% 会话数Sessions 19,934
› 11	小米Xiaomi Mimo V2.5 Pro · MIT	6.61% ±0.86%

综合排名Overall Rank — 名次区间Rank Spread 37 ↔ 42 确认成功Confirmed Success ▲11.12%±2.18% 赞扬/抱怨Praise vs Complaint ▲10.56%±2.49% 可控性Steerability ▲4.52%±1.69% Bash 恢复Bash Recovery ▲5.90%±1.51% 工具幻觉Tool Hallucination ▼0.92%±0.27% 会话数Sessions 44,611
› 12	MiniMaxMiniMax Minimax M3 · MiniMax Community License	6.81% ±0.84%

综合排名Overall Rank — 名次区间Rank Spread 37 ↔ 42 确认成功Confirmed Success ▲14.28%±2.28% 赞扬/抱怨Praise vs Complaint ▲13.18%±2.44% 可控性Steerability ▲6.99%±1.71% Bash 恢复Bash Recovery ▲0.51%±0.96% 工具幻觉Tool Hallucination ▼0.09%±0.19% 会话数Sessions 41,003
› 13	ThinkyThinky Inkling Small · Apache 2.0	9.60% ±1.35%

综合排名Overall Rank — 名次区间Rank Spread 42 ↔ 45 确认成功Confirmed Success ▲21.26%±4.01% 赞扬/抱怨Praise vs Complaint ▲20.52%±3.61% 可控性Steerability ▲11.51%±3.14% Bash 恢复Bash Recovery ▲5.72%±1.23% 工具幻觉Tool Hallucination ▼0.42%±0.32% 会话数Sessions 13,095
› 14	MistralMistral Mistral Medium 3.5 · Modified MIT	11.95% ±1.46%

综合排名Overall Rank — 名次区间Rank Spread 43 ↔ 45 确认成功Confirmed Success ▲16.36%±3.69% 赞扬/抱怨Praise vs Complaint ▲15.47%±3.98% 可控性Steerability ▲7.18%±3.25% Bash 恢复Bash Recovery ▲18.26%±2.70% 工具幻觉Tool Hallucination ▼2.48%±0.82% 会话数Sessions 10,858
› 15	UpstageUpstage Solar Pro 4 · Proprietary	16.25% ±1.87%

综合排名Overall Rank — 名次区间Rank Spread 46 ↔ 46 确认成功Confirmed Success ▲23.15%±4.88% 赞扬/抱怨Praise vs Complaint ▲20.12%±4.89% 可控性Steerability ▲10.33%±3.67% Bash 恢复Bash Recovery ▲25.73%±4.16% 工具幻觉Tool Hallucination ▼1.89%±0.74% 会话数Sessions 9,308

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