Arena 评分榜
AI 模型评分榜。
按 Arena 上的人类一对一盲测投票排名,附 95% 置信区间。
数据更新于
排行榜 · 41695% 区间
41NEW
glm-5.3-flash
1,474
zai · 32,925 票
42NEW
claude-opus-4-5-20251101-high-32k
1,474
anthropic · 37,509 票
43NEW
gpt-5.5-instant
1,473
openai · 27,158 票
44NEW

gemini-3-flash
1,473
google · 31,325 票
45NEW
claude-sonnet-4-6
1,472
anthropic · 70,931 票
46NEW
grok-4.20-beta-0309-reasoning
1,472
xai · 66,436 票
47NEW
grok-4.20-multi-agent-beta-0309
1,470
xai · 64,946 票
48NEW
claude-opus-4-5-20251101
1,470
anthropic · 72,899 票
49NEW
mimo-v2.5-pro
1,467
xiaomi · 75,721 票
50NEW
ernie-5.1
1,467
baidu · 39,468 票
51NEW
gpt-5.6-terra-xhigh
1,466
openai · 43,271 票
52NEW
grok-4.5
1,466
xai · 39,896 票
53NEW
grok-4.1-thinking
1,465
xai · 66,509 票
54NEW
glm-5.1
1,465
zai · 62,889 票
55NEW
gpt-5.4
1,465
openai · 67,870 票
56NEW
deepseek-v4-pro-high-20260813
1,465
deepseek · 9,800 票
57NEW

qwen3.5-max-preview
1,464
alibaba · 22,600 票
58NEW
claude-sonnet-5-high
1,461
anthropic · 49,595 票
59NEW
kimi-k2.6
1,461
moonshot · 40,013 票
60NEW

qwen3.6-max-preview
1,461
alibaba · 5,426 票
61NEW
gpt-6-sol-max
1,459
openai · 12,466 票
62NEW
grok-4.1
1,459
xai · 68,648 票
63NEW

gemini-3-flash (thinking-minimal)
1,458
google · 91,039 票
64NEW
deepseek-v4-pro
1,458
deepseek · 57,594 票
65NEW
glm-5
1,457
zai · 29,272 票
66NEW
claude-sonnet-4-5-20250929-high-32k
1,456
anthropic · 83,940 票
67NEW
dola-seed-2.0-pro
1,456
bytedance · 79,140 票
68NEW
gpt-5.1-high
1,456
openai · 41,651 票
69NEW

qwen3.7-plus
1,456
alibaba · 42,075 票
70NEW
claude-sonnet-4-5-20250929
1,455
anthropic · 82,688 票
71NEW
deepseek-v4-pro-high-preview
1,455
deepseek · 54,902 票
72NEW

gemini-3.5-flash-lite
1,455
google · 35,851 票
73NEW
grok-4.6-high
1,454
xai · 28,965 票
74NEW
hy3
1,453
tencent · 12,208 票
75NEW
gpt-5.6-luna-xhigh
1,453
openai · 41,109 票
76NEW

gemma-4-31b
1,452
google · 6,147 票
77NEW
mimo-v2.6-flash
1,451
xiaomi · 10,917 票
78NEW
gpt-5.3-chat-latest
1,450
openai · 34,345 票
79NEW
kimi-k2.5-thinking
1,450
moonshot · 74,759 票
80NEW
claude-opus-4-1-20250805-thinking-16k
1,450
anthropic · 49,542 票