Name: 智能体模型排行榜（对话任务模型2026-09-10排行榜）最好的 Agent 模型对比 |AITNT
Creator: LM Arena
License: https://creativecommons.org/licenses/by/4.0/
Keywords: 对话任务,AI 智能体模型排行榜,AI 模型排行

排名Rank ⇕	模型Model ⇕
› 🥇	GPT 6 Astra (Max) OpenAI · Proprietary	12.91% ±5.25%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 8 确认成功Confirmed Success ▲22.20%±7.09% 赞扬/抱怨Praise vs Complaint ▲45.41%±21.23% 可控性Steerability ▲10.10%±9.32% Bash 恢复Bash Recovery ▲6.73%±3.74% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 3,729
› 🥈	Claude Opus 5 (Max) Anthropic · Proprietary	11.81% ±3.49%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 7 确认成功Confirmed Success ▲16.68%±6.48% 赞扬/抱怨Praise vs Complaint ▲15.44%±13.56% 可控性Steerability ▲13.80%±6.60% Bash 恢复Bash Recovery ▲12.81%±1.89% 工具幻觉Tool Hallucination ▼0.31%±0.07% 会话数Sessions 8,186
› 🥉	Claude Opus 5 (High) Anthropic · Proprietary	10.59% ±3.37%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 10 确认成功Confirmed Success ▲8.57%±6.42% 赞扬/抱怨Praise vs Complaint ▲18.68%±13.12% 可控性Steerability ▲12.91%±5.95% Bash 恢复Bash Recovery ▲12.47%±2.15% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 10,322
› 4	Claude Fable 5.1 (Max) Anthropic · Proprietary	10.35% ±3.52%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 10 确认成功Confirmed Success ▲22.89%±5.41% 赞扬/抱怨Praise vs Complaint ▲23.17%±14.40% 可控性Steerability ▲4.66%±6.90% Bash 恢复Bash Recovery ▲9.99%±1.85% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 5,132
› 5	GPT 5.6 Sol (xHigh) OpenAI · Proprietary	7.54% ±2.69%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 13 确认成功Confirmed Success ▲8.66%±4.99% 赞扬/抱怨Praise vs Complaint ▲15.66%±10.30% 可控性Steerability ▲4.31%±5.15% Bash 恢复Bash Recovery ▲8.75%±1.62% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 13,218
› 6	Gemini 3.8 Flash (High) Google · Proprietary	7.19% ±3.77%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 20 确认成功Confirmed Success ▲11.70%±6.19% 赞扬/抱怨Praise vs Complaint ▲28.54%±13.11% 可控性Steerability ▲8.73%±11.31% Bash 恢复Bash Recovery ▲4.39%±2.56% 工具幻觉Tool Hallucination ▼0.08%±0.35% 会话数Sessions 2,863
› 7	Claude Fable 5 (High) Anthropic · Proprietary	7.06% ±2.83%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 14 确认成功Confirmed Success ▲10.36%±4.92% 赞扬/抱怨Praise vs Complaint ▲11.03%±9.26% 可控性Steerability ▲7.18%±4.57% Bash 恢复Bash Recovery ▲6.40%±7.72% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 15,893
› 8	GPT 5.5 (xHigh) OpenAI · Proprietary	5.14% ±2.14%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 21 确认成功Confirmed Success ▲1.40%±4.15% 赞扬/抱怨Praise vs Complaint ▲9.12%±7.83% 可控性Steerability ▲4.05%±3.86% Bash 恢复Bash Recovery ▲10.80%±3.19% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 22,329
› 9	Claude Opus 4.8 (High) Anthropic · Proprietary	5.09% ±2.68%

综合排名Overall Rank — 名次区间Rank Spread 2 ↔ 23 确认成功Confirmed Success ▲7.43%±4.94% 赞扬/抱怨Praise vs Complaint ▲4.88%±8.89% 可控性Steerability ▲7.21%±4.71% Bash 恢复Bash Recovery ▲6.09%±6.08% 工具幻觉Tool Hallucination ▼0.17%±0.67% 会话数Sessions 14,876
› 10	Qwen3.8 Max 阿里巴巴 · Proprietary	4.66% ±1.98%

综合排名Overall Rank — 名次区间Rank Spread 5 ↔ 22 确认成功Confirmed Success ▲10.84%±3.79% 赞扬/抱怨Praise vs Complaint ▲10.14%±7.34% 可控性Steerability ▲0.22%±3.91% Bash 恢复Bash Recovery ▲2.84%±2.32% 工具幻觉Tool Hallucination ▼0.72%±0.41% 会话数Sessions 10,071
› 11	Claude Sonnet 5 (High) Anthropic · Proprietary	4.36% ±3.20%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 29 确认成功Confirmed Success ▲0.93%±6.31% 赞扬/抱怨Praise vs Complaint ▲8.05%±11.00% 可控性Steerability ▲8.18%±6.23% Bash 恢复Bash Recovery ▲6.25%±5.48% 工具幻觉Tool Hallucination ▼0.25%±0.13% 会话数Sessions 12,797
› 12	Grok 4.5 SpaceXAI · Proprietary	4.29% ±2.03%

综合排名Overall Rank — 名次区间Rank Spread 5 ↔ 25 确认成功Confirmed Success ▲4.03%±4.26% 赞扬/抱怨Praise vs Complaint ▲2.32%±7.52% 可控性Steerability ▲5.06%±4.02% Bash 恢复Bash Recovery ▲9.72%±1.80% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 15,863
› 13	Kimi K3 (Max) 月之暗面 · Kimi K3 license	4.14% ±1.28%

综合排名Overall Rank — 名次区间Rank Spread 5 ↔ 21 确认成功Confirmed Success ▲14.30%±2.53% 赞扬/抱怨Praise vs Complaint ▲7.52%±4.61% 可控性Steerability ▲5.35%±2.84% Bash 恢复Bash Recovery ▲3.91%±1.34% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 44,042
› 14	DeepSeek V4 Pro (High) (0813) DeepSeek · MIT	2.14% ±1.70%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 30 确认成功Confirmed Success ▲6.76%±3.37% 赞扬/抱怨Praise vs Complaint ▲3.42%±6.21% 可控性Steerability ▲4.83%±3.58% Bash 恢复Bash Recovery ▲5.04%±1.34% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 15,324
› 15	GPT 5.5 OpenAI · Proprietary	2.07% ±1.76%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 30 确认成功Confirmed Success ▲0.46%±3.69% 赞扬/抱怨Praise vs Complaint ▲1.86%±6.19% 可控性Steerability ▲1.04%±3.26% Bash 恢复Bash Recovery ▲9.69%±2.15% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 28,503
› 16	Grok 4.6 (xHigh) SpaceXAI · Proprietary	1.88% ±2.14%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 31 确认成功Confirmed Success ▲5.89%±4.39% 赞扬/抱怨Praise vs Complaint ▲0.40%±7.48% 可控性Steerability ▲3.09%±4.60% Bash 恢复Bash Recovery ▲5.87%±1.78% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 9,108
› 17	Muse Spark 1.3 (Max) Meta · Proprietary	1.58% ±3.11%

综合排名Overall Rank — 名次区间Rank Spread 6 ↔ 36 确认成功Confirmed Success ▲9.09%±5.67% 赞扬/抱怨Praise vs Complaint ▲2.72%±10.26% 可控性Steerability ▲4.61%±8.95% Bash 恢复Bash Recovery ▲5.83%±2.90% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 3,439
› 18	GLM 5.2 (Max) 智谱 ZAI · MIT	1.31% ±1.43%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 31 确认成功Confirmed Success ▲2.98%±2.83% 赞扬/抱怨Praise vs Complaint ▲4.67%±4.79% 可控性Steerability ▲1.02%±2.66% Bash 恢复Bash Recovery ▲0.40%±2.56% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 31,030
› 19	Muse Spark 1.2 (xHigh) Meta · Proprietary	1.22% ±2.88%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 36 确认成功Confirmed Success ▲4.94%±5.37% 赞扬/抱怨Praise vs Complaint ▲3.52%±10.04% 可控性Steerability ▲2.82%±5.27% Bash 恢复Bash Recovery ▲0.11%±5.04% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 9,996
› 20	Hy4 preview 腾讯 · Apache 2.0	1.15% ±2.41%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 35 确认成功Confirmed Success ▲5.58%±4.94% 赞扬/抱怨Praise vs Complaint ▲0.28%±8.73% 可控性Steerability ▲5.08%±5.70% Bash 恢复Bash Recovery ▲5.38%±1.40% 工具幻觉Tool Hallucination ▼0.15%±0.15% 会话数Sessions 7,169
› 21	DeepSeek V4 Pro DeepSeek · MIT	1.06% ±2.33%

综合排名Overall Rank — 名次区间Rank Spread 8 ↔ 35 确认成功Confirmed Success ▲1.22%±5.10% 赞扬/抱怨Praise vs Complaint ▲0.79%±8.29% 可控性Steerability ▲1.38%±4.29% Bash 恢复Bash Recovery ▲5.76%±1.82% 工具幻觉Tool Hallucination ▼0.15%±0.16% 会话数Sessions 13,896
› 22	Claude Sonnet 4.6 Anthropic · Proprietary	1.00% ±2.60%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 36 确认成功Confirmed Success ▲7.21%±5.93% 赞扬/抱怨Praise vs Complaint ▲2.87%±8.90% 可控性Steerability ▲2.14%±5.00% Bash 恢复Bash Recovery ▲6.88%±2.66% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 10,825
› 23	GLM 5.3 Flash 智谱 ZAI · MIT	0.73% ±1.66%

综合排名Overall Rank — 名次区间Rank Spread 12 ↔ 33 确认成功Confirmed Success ▲12.05%±3.16% 赞扬/抱怨Praise vs Complaint ▲3.61%±5.88% 可控性Steerability ▲6.58%±3.67% Bash 恢复Bash Recovery ▲5.77%±2.06% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 11,214
› 24	GLM 5.3 (Max) 智谱 ZAI · MIT	0.46% ±1.24%

综合排名Overall Rank — 名次区间Rank Spread 13 ↔ 33 确认成功Confirmed Success ▲7.57%±2.44% 赞扬/抱怨Praise vs Complaint ▲1.86%±4.20% 可控性Steerability ▲4.91%±2.50% Bash 恢复Bash Recovery ▲2.57%±1.90% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 27,196
› 25	GPT 5.6 Luna (xHigh) OpenAI · Proprietary	0.24% ±2.08%

综合排名Overall Rank — 名次区间Rank Spread 12 ↔ 36 确认成功Confirmed Success ▲2.38%±4.11% 赞扬/抱怨Praise vs Complaint ▲0.71%±7.28% 可控性Steerability ▲1.86%±4.01% Bash 恢复Bash Recovery ▲4.40%±2.88% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 9,476
› 26	GPT 5.6 Terra (xHigh) OpenAI · Proprietary	0.20% ±2.30%

综合排名Overall Rank — 名次区间Rank Spread 11 ↔ 37 确认成功Confirmed Success ▲4.07%±4.86% 赞扬/抱怨Praise vs Complaint ▲1.83%±8.00% 可控性Steerability ▲0.70%±4.51% Bash 恢复Bash Recovery ▲7.26%±3.50% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 8,257
› 27	Muse Spark 1.1 Meta · Proprietary	0.20% ±1.37%

综合排名Overall Rank — 名次区间Rank Spread 13 ↔ 36 确认成功Confirmed Success ▲11.06%±2.53% 赞扬/抱怨Praise vs Complaint ▲5.02%±4.18% 可控性Steerability ▲5.60%±2.76% Bash 恢复Bash Recovery ▲1.77%±3.68% 工具幻觉Tool Hallucination ▼0.32%±0.07% 会话数Sessions 40,841
› 28	GPT 5.4 (High) OpenAI · Proprietary	0.48% ±1.72%

综合排名Overall Rank — 名次区间Rank Spread 13 ↔ 37 确认成功Confirmed Success ▲1.90%±3.81% 赞扬/抱怨Praise vs Complaint ▲8.10%±6.04% 可控性Steerability ▲0.45%±3.53% Bash 恢复Bash Recovery ▲6.85%±2.07% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 27,543
› 29	Deepseek V4 Flash (High) (20260731) DeepSeek · MIT	0.54% ±1.47%

综合排名Overall Rank — 名次区间Rank Spread 14 ↔ 37 确认成功Confirmed Success ▲1.50%±2.97% 赞扬/抱怨Praise vs Complaint ▲1.04%±5.03% 可控性Steerability ▲5.31%±3.42% Bash 恢复Bash Recovery ▲0.23%±1.30% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 25,970
› 30	Gemini 3.6 Flash (High) Google · Proprietary	0.57% ±2.22%

综合排名Overall Rank — 名次区间Rank Spread 13 ↔ 38 确认成功Confirmed Success ▲0.19%±4.66% 赞扬/抱怨Praise vs Complaint ▲0.01%±7.72% 可控性Steerability ▲1.20%±4.33% Bash 恢复Bash Recovery ▲4.10%±2.67% 工具幻觉Tool Hallucination ▼0.25%±0.15% 会话数Sessions 8,269
› 31	Qwen3.7 Max 阿里巴巴 · Proprietary	2.33% ±2.21%

综合排名Overall Rank — 名次区间Rank Spread 16 ↔ 40 确认成功Confirmed Success ▲5.45%±4.29% 赞扬/抱怨Praise vs Complaint ▲3.00%±6.54% 可控性Steerability ▲1.21%±3.60% Bash 恢复Bash Recovery ▲1.83%±6.33% 工具幻觉Tool Hallucination ▼0.15%±0.39% 会话数Sessions 15,817
› 32	Qwen3.8 Flash Next 阿里巴巴 · Qwen-community-1.0	2.43% ±1.74%

综合排名Overall Rank — 名次区间Rank Spread 18 ↔ 40 确认成功Confirmed Success ▲8.50%±3.56% 赞扬/抱怨Praise vs Complaint ▲4.05%±5.76% 可控性Steerability ▲9.07%±4.28% Bash 恢复Bash Recovery ▲6.80%±1.46% 工具幻觉Tool Hallucination ▼0.74%±0.27% 会话数Sessions 11,751
› 33	Qwen 3.8 27B 阿里巴巴 · Apache 2.0	2.97% ±1.73%

综合排名Overall Rank — 名次区间Rank Spread 20 ↔ 40 确认成功Confirmed Success ▲4.46%±3.47% 赞扬/抱怨Praise vs Complaint ▲4.85%±5.93% 可控性Steerability ▲3.52%±3.76% Bash 恢复Bash Recovery ▲10.94%±2.27% 工具幻觉Tool Hallucination ▼0.01%±0.18% 会话数Sessions 11,662
› 34	Mimo V2.5 Pro 小米 · MIT	3.39% ±2.42%

综合排名Overall Rank — 名次区间Rank Spread 20 ↔ 41 确认成功Confirmed Success ▲5.81%±4.48% 赞扬/抱怨Praise vs Complaint ▲1.93%±7.38% 可控性Steerability ▲2.04%±3.94% Bash 恢复Bash Recovery ▲6.86%±6.85% 工具幻觉Tool Hallucination ▼0.33%±0.47% 会话数Sessions 15,762
› 35	Qwen3.7 Plus 阿里巴巴 · Proprietary	3.79% ±2.42%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 41 确认成功Confirmed Success ▲5.15%±5.47% 赞扬/抱怨Praise vs Complaint ▲11.18%±7.80% 可控性Steerability ▲3.48%±4.94% Bash 恢复Bash Recovery ▲1.81%±4.39% 工具幻觉Tool Hallucination ▼0.93%±0.86% 会话数Sessions 8,815
› 36	Inkling Small Thinky · Apache 2.0	3.88% ±3.16%

综合排名Overall Rank — 名次区间Rank Spread 18 ↔ 41 确认成功Confirmed Success ▲15.61%±7.52% 赞扬/抱怨Praise vs Complaint ▲10.84%±10.91% 可控性Steerability ▲3.47%±6.97% Bash 恢复Bash Recovery ▲4.09%±2.66% 工具幻觉Tool Hallucination ▼0.52%±0.97% 会话数Sessions 4,696
› 37	Gemini 3.1 Pro Preview Google · Proprietary	3.90% ±1.95%

综合排名Overall Rank — 名次区间Rank Spread 27 ↔ 41 确认成功Confirmed Success ▲2.97%±3.82% 赞扬/抱怨Praise vs Complaint ▲2.01%±5.86% 可控性Steerability ▲0.78%±3.06% Bash 恢复Bash Recovery ▲19.64%±5.37% 工具幻觉Tool Hallucination ▼0.33%±0.06% 会话数Sessions 31,537
› 38	Hy3 腾讯 · Apache 2.0	5.30% ±2.31%

综合排名Overall Rank — 名次区间Rank Spread 31 ↔ 42 确认成功Confirmed Success ▲14.55%±5.18% 赞扬/抱怨Praise vs Complaint ▲5.29%±7.82% 可控性Steerability ▲0.73%±4.45% Bash 恢复Bash Recovery ▲3.23%±3.51% 工具幻觉Tool Hallucination ▼2.68%±1.21% 会话数Sessions 10,672
› 39	Gemini 3.5 Flash Lite Google · Proprietary	5.67% ±3.05%

综合排名Overall Rank — 名次区间Rank Spread 30 ↔ 42 确认成功Confirmed Success ▲13.98%±6.25% 赞扬/抱怨Praise vs Complaint ▲1.46%±9.39% 可控性Steerability ▲0.44%±5.43% Bash 恢复Bash Recovery ▲11.32%±6.39% 工具幻觉Tool Hallucination ▼1.17%±1.79% 会话数Sessions 9,541
› 40	Minimax M3 MiniMax · MiniMax Community License	5.95% ±1.81%

综合排名Overall Rank — 名次区间Rank Spread 31 ↔ 42 确认成功Confirmed Success ▲12.74%±4.49% 赞扬/抱怨Praise vs Complaint ▲13.90%±5.76% 可控性Steerability ▲3.85%±3.55% Bash 恢复Bash Recovery ▲0.91%±2.63% 工具幻觉Tool Hallucination ▼0.15%±0.43% 会话数Sessions 16,086
› 41	Inkling Thinky · Apache 2.0	7.03% ±2.10%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 43 确认成功Confirmed Success ▲16.51%±5.45% 赞扬/抱怨Praise vs Complaint ▲19.65%±6.43% 可控性Steerability ▲0.60%±4.53% Bash 恢复Bash Recovery ▲0.57%±3.37% 工具幻觉Tool Hallucination ▼0.16%±0.43% 会话数Sessions 16,315
› 42	Mistral Medium 3.5 Mistral · Modified MIT	10.01% ±2.84%

综合排名Overall Rank — 名次区间Rank Spread 38 ↔ 43 确认成功Confirmed Success ▲13.37%±6.04% 赞扬/抱怨Praise vs Complaint ▲21.57%±6.82% 可控性Steerability ▲6.50%±5.74% Bash 恢复Bash Recovery ▲18.15%±8.08% 工具幻觉Tool Hallucination ▼3.47%±2.81% 会话数Sessions 4,006
› 43	Minimax M2.7 MiniMax · Modified MIT	11.27% ±2.48%

综合排名Overall Rank — 名次区间Rank Spread 41 ↔ 43 确认成功Confirmed Success ▲19.00%±5.04% 赞扬/抱怨Praise vs Complaint ▲16.92%±6.43% 可控性Steerability ▲1.88%±4.40% Bash 恢复Bash Recovery ▲18.69%±7.39% 工具幻觉Tool Hallucination ▼0.17%±0.15% 会话数Sessions 15,481

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