Valumigo

Design Arena · 游戏开发

定期获取并展示公开的基准测试资料(LMArena 用户投票排名、Epoch AI 测试分数、OpenRouter 使用量排名)。每个榜单均标注数据发布日期和获取日期。这些分数并非本站自行评定。

LMArena 分数来自用户并排查看两个模型的回答,再投票选出更好的一方(Elo 方式)。如果分数差异在置信区间内,可视为水平基本相近。

如何解读此排行榜

衡量什么: 这是 Design Arena 中,人们比较根据相同请求生成的简单游戏并投票得出的排名。

如何解读分数: 采用 Elo 方式计算相对得分,得分越高,表示人们更经常偏好该模型的结果。同时显示胜率。

注意事项: 主要评测简短的单个游戏,因此可能与大型项目开发能力不同。

本次排名的解读要点

  • 最高分为 Claude Opus 5.5 的 1479。
  • 前 10 个模型中,Anthropic 的模型有 4 个,数量最多。

游戏开发 Design Arena

发布日期:2026-10-05 · 获取日期:2026-10-05

125013001350140014501500
AnthropicClaude Opus 5.5
1479
AnthropicClaude Fable 5.1
1394
Moonshot AIKimi K3
1379
MetaMuse Spark 1.3 Max
1358
AnthropicClaude Opus 5
1350
AnthropicClaude Fable 5
1349
MetaMuse Spark 1.3 (xhigh)
1343
GoogleGemini 3.7 Flash
1335
ZGLM-5.3
1334
AlibabaQwen3.8 Max
1329
XiaomiMiMo-V2.6-Pro
1323
GoogleGemini 3.8 Flash
1320
MetaMuse Spark 1.2
1314
OGPT-5.5
1308
xGrok 4.6
1307
xGrok 4.7
1304
AnthropicClaude Sonnet 5
1296
ZGLM-5.3-Flash
1294
AnthropicClaude Opus 4.6 (Thinking)
1290
AnthropicClaude Opus 4.6
1285

图表仅显示同一模型各推理设置(high、max 等)中排名最高的一项。表格列出所有设置各自的排名。

查看表格(前 60 名)
排名模型开发方分数胜率
1#1AnthropicClaude Opus 5.5Anthropic147973.6%
2#2AnthropicClaude Fable 5.1Anthropic139465.6%
3#3Moonshot AIKimi K3Moonshot AI137961.8%
4#4MetaMuse Spark 1.3 MaxMeta135856.2%
5#5AnthropicClaude Opus 5Anthropic135059.1%
6#6AnthropicClaude Fable 5Anthropic134960.8%
7#7MetaMuse Spark 1.3 (xhigh)Meta134354.6%
8#8GoogleGemini 3.7 FlashGoogle133554.2%
9#9ZGLM-5.3Z.ai133454.6%
10#10AlibabaQwen3.8 MaxAlibaba132953.3%
11#11XiaomiMiMo-V2.6-ProXiaomi132353%
12#12GoogleGemini 3.8 FlashGoogle132051.3%
13#13MetaMuse Spark 1.2Meta131450.4%
14#14OGPT-5.5OpenAI130858.1%
15#15xGrok 4.6xAI130754.2%
16#16xGrok 4.7xAI130451.5%
17#17AnthropicClaude Sonnet 5Anthropic129653%
18#18ZGLM-5.3-FlashZ.ai129447.2%
19#19AnthropicClaude Opus 4.6 (Thinking)Anthropic129061%
20#20AnthropicClaude Opus 4.6Anthropic128560.5%
21#21MetaMuse Spark 1.1Meta128548.1%
22#22ZGLM 5.2Z.ai128149.5%
23#23AlibabaQwen3.7 MaxAlibaba128053.6%
24#24AnthropicClaude Opus 4.8Anthropic127553.2%
25#25ZGLM 5.1Z.ai127555.6%
26#26GoogleGemini 3.6 FlashGoogle127551.1%
27#27xGrok 4.5xAI127549.3%
28#28GoogleGemini 3.5 FlashGoogle127453%
29#29AlibabaQwen3.7 PlusAlibaba127349.7%
30#30AnthropicClaude Sonnet 4.6Anthropic127258.8%
31#31XiaomiMiMo-V2.5-ProXiaomi127253.4%
32#32Moonshot AIKimi K2.6Moonshot AI126354.9%
33#33ZGLM 5 TurboZ.ai126053.2%
34#34XiaomiMiMo-V2.5Xiaomi125752.6%
35#35OGPT-5.4 (Medium)OpenAI125157.6%
36#36ZGLM 5Z.ai124757.4%
37#37AnthropicClaude Opus 4.5Anthropic124259.3%
38#38DeepSeekDeepSeek-V4-ProDeepSeek124252.3%
39#39NNex N2 ProNex Agi123849%
40#40Moonshot AIKimi K2.7 CodeMoonshot AI123347.5%
41#41ZGLM 5V TurboZ.ai123150.7%
42#42MiniMaxMiniMax M3MiniMax122546%
43#43AlibabaQwen3.6 PlusAlibaba122449.1%
44#44MiniMaxMiniMax M2.7MiniMax122350%
45#45DeepSeekDeepSeek-V4-Flash-0731DeepSeek122144.6%
46#46Moonshot AIKimi K2.5 (Thinking)Moonshot AI121953.4%
47#47GoogleGemini 3.1 Pro PreviewGoogle121552.1%
48#48DeepSeekDeepSeek-V4-FlashDeepSeek121050.2%
49#49GoogleGemini 3 Pro PreviewGoogle120960.9%
50#50OGPT-5.2 (Low)OpenAI120856%
51#51OGPT-5.4 (None)OpenAI120753.6%
52#52xGrok 4.20 Beta (Reasoning)xAI120550.1%
53#53OGPT-5.4 (Low)OpenAI120452.8%
54#54AnthropicClaude 3.7 SonnetAnthropic120354.3%
55#55ZGLM 4.7Z.ai120155.2%
56#56OGPT-5.2 (High)OpenAI119956.7%
57#57OGPT-5 (High)OpenAI119859.3%
58#58XiaomiMiMo-V2-FlashXiaomi119748.6%
59#59OGPT-5.2 (Medium)OpenAI119652.8%
60#60OGPT-5.1 (High)OpenAI119155.9%

Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings). Licensed under CC BY 4.0.

数据:LMArena leaderboard dataset(CC BY 4.0)。修改:各榜单仅显示前 50 名,分数四舍五入。 https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset公司标志为各公司的商标,仅用于帮助区分(图标:Simple Icons)。数据:Epoch AI — Capabilities & benchmarking(CC BY 4.0)。 https://epoch.ai/benchmarksSource: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings). Source: OpenRouter evals (openrouter.ai) via OpenRouter (openrouter.ai/rankings). Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings). Licensed under CC BY 4.0. Source: OpenRouter (openrouter.ai/rankings), as of 2026-10-05. Licensed under CC BY 4.0. https://openrouter.ai/rankings