Name: 智能体 AI 排行榜 - 最佳 Agent 模型对比 (2026-07-19 历史榜单)
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 ⇕	模型Model ⇕
› 🥇	Claude Fable 5 (High) Anthropic · Proprietary	13.22% ±1.98%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 4 确认成功Confirmed Success ▲11.71%±3.77% 赞扬/抱怨Praise vs Complaint ▲24.97%±7.41% 可控性Steerability ▲15.64%±3.76% Bash 恢复Bash Recovery ▲12.67%±1.31% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 23,509
› 🥈	Claude Opus 4.8 (Thinking) Anthropic · Proprietary	10.04% ±1.39%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 8 确认成功Confirmed Success ▲9.11%±2.62% 赞扬/抱怨Praise vs Complaint ▲19.58%±5.05% 可控性Steerability ▲11.00%±2.58% Bash 恢复Bash Recovery ▲10.59%±1.05% 工具幻觉Tool Hallucination ▼0.09%±1.12% 会话数Sessions 34,104
› 🥉	GPT 5.6 Sol (xHigh) OpenAI · Proprietary	9.86% ±1.66%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 9 确认成功Confirmed Success ▲7.64%±3.30% 赞扬/抱怨Praise vs Complaint ▲21.97%±6.41% 可控性Steerability ▲9.95%±2.77% Bash 恢复Bash Recovery ▲8.64%±1.22% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 15,457
› 4	Kimi K3 月之暗面 · Proprietary	9.62% ±1.87%

综合排名Overall Rank — 名次区间Rank Spread 1 ↔ 10 确认成功Confirmed Success ▲14.42%±3.44% 赞扬/抱怨Praise vs Complaint ▲20.62%±6.90% 可控性Steerability ▲5.58%±4.07% Bash 恢复Bash Recovery ▲6.41%±1.52% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 8,344
› 5	Claude Sonnet 5 (High) Anthropic · Proprietary	9.10% ±1.87%

综合排名Overall Rank — 名次区间Rank Spread 2 ↔ 12 确认成功Confirmed Success ▲8.70%±3.64% 赞扬/抱怨Praise vs Complaint ▲17.59%±7.04% 可控性Steerability ▲7.46%±3.60% Bash 恢复Bash Recovery ▲10.78%±0.88% 工具幻觉Tool Hallucination ▼0.95%±0.13% 会话数Sessions 24,310
› 6	GPT 5.5 (xHigh) OpenAI · Proprietary	8.80% ±0.86%

综合排名Overall Rank — 名次区间Rank Spread 2 ↔ 9 确认成功Confirmed Success ▲7.21%±1.76% 赞扬/抱怨Praise vs Complaint ▲11.69%±3.09% 可控性Steerability ▲9.22%±1.63% Bash 恢复Bash Recovery ▲14.78%±0.72% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 40,135
› 7	Claude Opus 4.7 (Thinking) Anthropic · Proprietary	8.30% ±1.25%

综合排名Overall Rank — 名次区间Rank Spread 2 ↔ 13 确认成功Confirmed Success ▲5.93%±2.56% 赞扬/抱怨Praise vs Complaint ▲12.39%±4.40% 可控性Steerability ▲9.72%±2.37% Bash 恢复Bash Recovery ▲12.51%±1.14% 工具幻觉Tool Hallucination ▼0.98%±0.14% 会话数Sessions 35,081
› 8	Claude Opus 4.7 Anthropic · Proprietary	8.07% ±1.25%

综合排名Overall Rank — 名次区间Rank Spread 2 ↔ 13 确认成功Confirmed Success ▲5.84%±2.56% 赞扬/抱怨Praise vs Complaint ▲12.89%±4.38% 可控性Steerability ▲9.87%±2.32% Bash 恢复Bash Recovery ▲10.72%±1.54% 工具幻觉Tool Hallucination ▼1.03%±0.13% 会话数Sessions 35,614
› 9	GPT 5.5 (High) OpenAI · Proprietary	7.73% ±0.79%

综合排名Overall Rank — 名次区间Rank Spread 3 ↔ 13 确认成功Confirmed Success ▲6.51%±1.57% 赞扬/抱怨Praise vs Complaint ▲10.04%±2.84% 可控性Steerability ▲9.32%±1.43% Bash 恢复Bash Recovery ▲11.70%±1.10% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 65,314
› 10	Claude Opus 4.6 Anthropic · Proprietary	6.65% ±1.24%

综合排名Overall Rank — 名次区间Rank Spread 5 ↔ 15 确认成功Confirmed Success ▲3.09%±2.65% 赞扬/抱怨Praise vs Complaint ▲10.22%±4.21% 可控性Steerability ▲7.67%±2.30% Bash 恢复Bash Recovery ▲11.19%±1.34% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 34,797
› 11	GLM 5.2 (Max) 智谱 ZAI · MIT	6.47% ±1.02%

综合排名Overall Rank — 名次区间Rank Spread 6 ↔ 15 确认成功Confirmed Success ▲8.53%±2.02% 赞扬/抱怨Praise vs Complaint ▲12.29%±3.67% 可控性Steerability ▲5.59%±1.82% Bash 恢复Bash Recovery ▲4.85%±1.17% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 37,177
› 12	GPT 5.5 OpenAI · Proprietary	6.43% ±0.76%

综合排名Overall Rank — 名次区间Rank Spread 7 ↔ 15 确认成功Confirmed Success ▲4.87%±1.54% 赞扬/抱怨Praise vs Complaint ▲7.50%±2.69% 可控性Steerability ▲7.40%±1.36% Bash 恢复Bash Recovery ▲11.28%±0.90% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 66,256
› 13	Grok 4.5 SpaceXAI · Proprietary	6.42% ±1.28%

综合排名Overall Rank — 名次区间Rank Spread 6 ↔ 15 确认成功Confirmed Success ▲6.21%±2.67% 赞扬/抱怨Praise vs Complaint ▲9.01%±4.72% 可控性Steerability ▲4.95%±2.21% Bash 恢复Bash Recovery ▲10.86%±1.01% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 20,900
› 14	GPT 5.4 (High) OpenAI · Proprietary	5.90% ±0.75%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 15 确认成功Confirmed Success ▲6.88%±1.56% 赞扬/抱怨Praise vs Complaint ▲3.56%±2.64% 可控性Steerability ▲8.41%±1.44% Bash 恢复Bash Recovery ▲9.58%±0.92% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 65,579
› 15	Claude Opus 4.8 Anthropic · Proprietary	4.03% ±1.65%

综合排名Overall Rank — 名次区间Rank Spread 10 ↔ 17 确认成功Confirmed Success ▲7.48%±2.73% 赞扬/抱怨Praise vs Complaint ▲12.74%±4.86% 可控性Steerability ▲9.06%±2.64% Bash 恢复Bash Recovery ▲9.76%±1.39% 工具幻觉Tool Hallucination ▼18.89%±4.50% 会话数Sessions 32,151
› 16	Claude Sonnet 4.6 Anthropic · Proprietary	3.15% ±1.15%

综合排名Overall Rank — 名次区间Rank Spread 15 ↔ 18 确认成功Confirmed Success ▲0.23%±2.63% 赞扬/抱怨Praise vs Complaint ▲1.10%±3.78% 可控性Steerability ▲2.31%±2.19% Bash 恢复Bash Recovery ▲11.53%±1.39% 工具幻觉Tool Hallucination ▼1.05%±0.12% 会话数Sessions 35,582
› 17	GLM 5.1 智谱 ZAI · MIT	1.66% ±0.79%

综合排名Overall Rank — 名次区间Rank Spread 15 ↔ 20 确认成功Confirmed Success ▲1.66%±1.74% 赞扬/抱怨Praise vs Complaint ▲1.25%±2.71% 可控性Steerability ▲0.68%±1.55% Bash 恢复Bash Recovery ▲3.61%±0.91% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 56,470
› 18	Muse Spark 1.1 Meta · Proprietary	1.22% ±0.95%

综合排名Overall Rank — 名次区间Rank Spread 16 ↔ 22 确认成功Confirmed Success ▲4.97%±2.15% 赞扬/抱怨Praise vs Complaint ▲2.32%±2.96% 可控性Steerability ▲4.01%±1.80% Bash 恢复Bash Recovery ▲6.39%±1.67% 工具幻觉Tool Hallucination ▼1.06%±0.12% 会话数Sessions 24,882
› 19	Qwen3.7 Max 阿里巴巴 · Proprietary	0.33% ±1.10%

综合排名Overall Rank — 名次区间Rank Spread 17 ↔ 25 确认成功Confirmed Success ▲1.91%±2.71% 赞扬/抱怨Praise vs Complaint ▲4.71%±3.60% 可控性Steerability ▲0.62%±2.09% Bash 恢复Bash Recovery ▲7.09%±1.45% 工具幻觉Tool Hallucination ▼0.55%±0.27% 会话数Sessions 14,960
› 20	Gemini 3.1 Pro Preview Google · Proprietary	0.19% ±0.68%

综合排名Overall Rank — 名次区间Rank Spread 18 ↔ 25 确认成功Confirmed Success ▲2.13%±1.50% 赞扬/抱怨Praise vs Complaint ▲1.08%±2.23% 可控性Steerability ▲3.26%±1.23% Bash 恢复Bash Recovery ▲8.41%±1.16% 工具幻觉Tool Hallucination ▼1.01%±0.13% 会话数Sessions 66,608
› 21	Gemini 3.5 Flash (High) Google · Proprietary	0.67% ±0.80%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 25 确认成功Confirmed Success ▲3.09%±1.76% 赞扬/抱怨Praise vs Complaint ▲3.44%±2.47% 可控性Steerability ▲0.49%±1.45% Bash 恢复Bash Recovery ▲1.78%±1.51% 工具幻觉Tool Hallucination ▼1.72%±0.37% 会话数Sessions 45,992
› 22	Kimi K2.7 Code 月之暗面 · Modified MIT	0.74% ±1.71%

综合排名Overall Rank — 名次区间Rank Spread 17 ↔ 27 确认成功Confirmed Success ▲4.30%±3.47% 赞扬/抱怨Praise vs Complaint ▲2.04%±6.08% 可控性Steerability ▲8.21%±3.21% Bash 恢复Bash Recovery ▲2.92%±2.79% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 10,028
› 23	Qwen3.7 Plus 阿里巴巴 · Proprietary	0.83% ±1.23%

综合排名Overall Rank — 名次区间Rank Spread 18 ↔ 27 确认成功Confirmed Success ▲1.59%±3.07% 赞扬/抱怨Praise vs Complaint ▲6.37%±3.78% 可控性Steerability ▲1.06%±2.50% Bash 恢复Bash Recovery ▲4.94%±1.95% 工具幻觉Tool Hallucination ▼0.05%±0.52% 会话数Sessions 12,551
› 24	DeepSeek V4 Pro DeepSeek · MIT	0.84% ±1.09%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 27 确认成功Confirmed Success ▲3.42%±2.76% 赞扬/抱怨Praise vs Complaint ▲5.50%±3.51% 可控性Steerability ▲1.45%±2.12% Bash 恢复Bash Recovery ▲5.63%±1.02% 工具幻觉Tool Hallucination ▼0.56%±0.24% 会话数Sessions 15,455
› 25	Kimi K2.6 月之暗面 · Modified MIT	2.26% ±1.75%

综合排名Overall Rank — 名次区间Rank Spread 19 ↔ 28 确认成功Confirmed Success ▲2.03%±3.50% 赞扬/抱怨Praise vs Complaint ▲2.08%±5.51% 可控性Steerability ▲6.26%±3.20% Bash 恢复Bash Recovery ▲6.06%±3.91% 工具幻觉Tool Hallucination ▼1.08%±0.12% 会话数Sessions 10,092
› 26	Minimax M3 MiniMax · MiniMax Community License	2.74% ±1.08%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 28 确认成功Confirmed Success ▲7.09%±2.81% 赞扬/抱怨Praise vs Complaint ▲9.27%±3.42% 可控性Steerability ▲4.25%±2.21% Bash 恢复Bash Recovery ▲6.12%±0.95% 工具幻觉Tool Hallucination ▼0.77%±0.21% 会话数Sessions 14,976
› 27	Mimo V2.5 Pro 小米 · MIT	2.98% ±1.13%

综合排名Overall Rank — 名次区间Rank Spread 22 ↔ 29 确认成功Confirmed Success ▲6.30%±2.84% 赞扬/抱怨Praise vs Complaint ▲8.88%±3.51% 可控性Steerability ▲2.50%±2.21% Bash 恢复Bash Recovery ▲2.53%±1.68% 工具幻觉Tool Hallucination ▼0.22%±0.32% 会话数Sessions 15,454
› 28	DeepSeek V4 Flash DeepSeek · MIT	3.78% ±1.07%

综合排名Overall Rank — 名次区间Rank Spread 25 ↔ 30 确认成功Confirmed Success ▲7.23%±2.89% 赞扬/抱怨Praise vs Complaint ▲10.83%±3.28% 可控性Steerability ▲3.36%±2.12% Bash 恢复Bash Recovery ▲3.41%±1.16% 工具幻觉Tool Hallucination ▼0.90%±0.42% 会话数Sessions 14,991
› 29	Inkling Thinky · Apache 2.0	5.70% ±1.60%

综合排名Overall Rank — 名次区间Rank Spread 27 ↔ 32 确认成功Confirmed Success ▲5.61%±4.17% 赞扬/抱怨Praise vs Complaint ▲17.97%±4.56% 可控性Steerability ▲10.42%±3.69% Bash 恢复Bash Recovery ▲6.41%±1.92% 工具幻觉Tool Hallucination ▼0.92%±0.58% 会话数Sessions 7,595
› 30	Gemini 3.5 Flash (Medium) Google · Proprietary	6.54% ±1.72%

综合排名Overall Rank — 名次区间Rank Spread 28 ↔ 33 确认成功Confirmed Success ▲13.06%±4.15% 赞扬/抱怨Praise vs Complaint ▲7.71%±5.06% 可控性Steerability ▲8.63%±3.29% Bash 恢复Bash Recovery ▲3.88%±3.51% 工具幻觉Tool Hallucination ▼0.58%±0.53% 会话数Sessions 8,380
› 31	Grok Build 0.1 SpaceXAI · Proprietary	7.56% ±0.81%

综合排名Overall Rank — 名次区间Rank Spread 29 ↔ 33 确认成功Confirmed Success ▲3.91%±1.76% 赞扬/抱怨Praise vs Complaint ▲11.39%±2.41% 可控性Steerability ▲11.00%±1.59% Bash 恢复Bash Recovery ▲11.94%±1.83% 工具幻觉Tool Hallucination ▼0.44%±0.13% 会话数Sessions 58,041
› 32	Grok 4.3 (High) SpaceXAI · Proprietary	7.85% ±0.81%

综合排名Overall Rank — 名次区间Rank Spread 29 ↔ 33 确认成功Confirmed Success ▲8.82%±1.73% 赞扬/抱怨Praise vs Complaint ▲13.97%±2.01% 可控性Steerability ▲5.96%±1.31% Bash 恢复Bash Recovery ▲11.24%±2.44% 工具幻觉Tool Hallucination ▼0.76%±0.13% 会话数Sessions 46,801
› 33	Gemini 3 Flash Google · Proprietary	8.33% ±0.75%

综合排名Overall Rank — 名次区间Rank Spread 30 ↔ 33 确认成功Confirmed Success ▲8.29%±1.58% 赞扬/抱怨Praise vs Complaint ▲12.10%±1.90% 可控性Steerability ▲4.34%±1.24% Bash 恢复Bash Recovery ▲16.93%±1.99% 工具幻觉Tool Hallucination ▼0.01%±1.04% 会话数Sessions 67,321
› 34	Minimax M2.7 MiniMax · Modified MIT	11.49% ±1.36%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 36 确认成功Confirmed Success ▲16.31%±3.20% 赞扬/抱怨Praise vs Complaint ▲14.20%±3.88% 可控性Steerability ▲16.01%±2.50% Bash 恢复Bash Recovery ▲11.83%±3.17% 工具幻觉Tool Hallucination ▼0.91%±0.18% 会话数Sessions 15,150
› 35	Nemotron 3 Ultra NVIDIA · OpenMDW-1.1	12.90% ±2.43%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 37 确认成功Confirmed Success ▲14.43%±5.14% 赞扬/抱怨Praise vs Complaint ▲10.80%±7.44% 可控性Steerability ▲19.21%±4.85% Bash 恢复Bash Recovery ▲19.61%±5.56% 工具幻觉Tool Hallucination ▼0.46%±0.69% 会话数Sessions 10,137
› 36	Gemma 4 31B Google · Apache 2.0	13.97% ±1.58%

综合排名Overall Rank — 名次区间Rank Spread 34 ↔ 37 确认成功Confirmed Success ▲2.13%±1.72% 赞扬/抱怨Praise vs Complaint ▲3.88%±2.60% 可控性Steerability ▲5.06%±1.51% Bash 恢复Bash Recovery ▲33.43%±5.08% 工具幻觉Tool Hallucination ▼25.33%±5.02% 会话数Sessions 54,361
› 37	Grok 4.3 SpaceXAI · Proprietary	14.68% ±1.03%

综合排名Overall Rank — 名次区间Rank Spread 35 ↔ 37 确认成功Confirmed Success ▲10.35%±1.62% 赞扬/抱怨Praise vs Complaint ▲15.32%±1.87% 可控性Steerability ▲6.93%±1.24% Bash 恢复Bash Recovery ▲41.67%±4.23% 工具幻觉Tool Hallucination ▼0.90%±0.13% 会话数Sessions 66,704

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