跳转到内容

实时模型榜

汇总持续更新的公开评测,保留每个模型的来源、原始成绩与覆盖情况。

实时汇总

5 个独立来源 · 6 张榜单 · 检查于 2026/09/16 01:05 · 内容更新于 2026/09/15 15:35

评分方法与来源说明

色块表示该模型被哪些榜单收录

排名 模型 发布时间 来源覆盖 共识分
01 Claude Fable 5.1 Anthropic 2026-09-01
4/6
98.0/100 证据 →
02 GPT-6 Astra OpenAI 2026-09-03
5/6
95.3/100 证据 →
03 Claude Opus 5 Anthropic 2026-07-24
5/6
93.4/100 证据 →
04 Claude Fable 5 Anthropic 2026-06-09
5/6
92.9/100 证据 →
05 GPT-5.6 Sol OpenAI 2026-07-09
5/6
87.1/100 证据 →
06 Gemini 3.8 Flash Google 2026-09-02
4/6
84.3/100 证据 →
07 Kimi K3 Moonshot AI 2026-07-16
5/6
81.5/100 证据 →
08 Claude Opus 4.7 Anthropic 2026-04-16
5/6
80.5/100 证据 →
09 GPT-5.5 OpenAI 2026-04-23
6/6
78.2/100 证据 →
10 GPT-5.3 Codex OpenAI 2026-02-05
3/6
76.7/100 证据 →
11 Claude Opus 4.6 Anthropic 2026-02-05
5/6
75.5/100 证据 →
12 Claude Opus 4.8 Anthropic 2026-05-28
5/6
74.1/100 证据 →
13 Gemini 3.7 Flash Google 2026-08-13
4/6
73.2/100 证据 →
14 GPT-5.2-Codex OpenAI 2025-12-18
2/6
72.6/100 证据 →
15 GPT-5.6 Terra OpenAI 2026-07-09
4/6
72.4/100 证据 →
16 GLM-5.3 Z.ai 2026-08-14
4/6
71.7/100 证据 →
17 DeepSeek V4 Pro 0813 DeepSeek 2026-08-13
2/6
71.2/100 证据 →
18 Grok 4.6 xAI 2026-08-12
5/6
70.5/100 证据 →
19 Qwen 3.8 Max Alibaba 2026-08-02
4/6
67.8/100 证据 →
20 GPT-5.4 OpenAI 2026-03-05
6/6
65.9/100 证据 →
21 DeepSeek V4 Flash 0731 DeepSeek 2026-07-31
2/6
63.6/100 证据 →
22 GPT-5.1-Codex OpenAI 2025-11-19
2/6
62.0/100 证据 →
23 Muse Spark 1.1 Meta 2026-07-09
5/6
61.8/100 证据 →
24 Muse Spark 1.2 Meta 2026-08-05
4/6
61.4/100 证据 →
25 Gemini 3 Pro Google 2025-11-18
5/6
61.1/100 证据 →
26 GPT-5.2 OpenAI 2025-12-11
4/6
60.8/100 证据 →
27 Gemini 3 Flash Google 2025-12-17
4/6
55.7/100 证据 →
28 GPT-5.6 Luna OpenAI 2026-07-09
4/6
55.5/100 证据 →
29 Gemini 3.1 Pro Google 2026-02-19
6/6
55.3/100 证据 →
30 Claude Sonnet 5 Anthropic 2026-06-30
5/6
54.7/100 证据 →