Valumigo

Artificial Analysis · 编程指数

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 Artificial Analysis 汇总编程相关评测得出的编程指数。

如何解读分数： 指数越高，表示在代码编写和修改评测中表现越好。

注意事项： 这是交互式编程评测的结果，可能与处理整个代码仓库的智能体任务能力有所不同。请同时参考我们的编程实测记录。

本次排名的解读要点

  • 最高分为 Claude Fable 5.1 (Max, Default Fallback) 的 81.6。
  • 第 1 名与第 5 名的差距为 4.8。
  • 前 10 个模型中，Anthropic 的模型有 3 个，数量最多。

编程指数 Artificial Analysis

发布日期：2026-10-05 · 获取日期：2026-10-05

AnthropicClaude Fable 5.1 (Max, Default Fallback)
81.6
OGPT-5.6 Sol (Xhigh)
78.3
AnthropicClaude Opus 5 (Max)
78.0
OGPT-6 Astra (max)
76.9
xGrok 4.6 (High)
76.8
OGPT-5.6 Terra (Max)
76.7
AnthropicClaude Fable 5 (Max, Opus 4.8 Fallback)
76.5
GoogleGemini 3.8 Flash (High)
76.3
Moonshot AIKimi K3 (Max)
76.2
AlibabaQwen3.8 Max (0902)
76.2
GoogleGemini 3.7 Flash (High)
76.1
OGPT-5.5 (Xhigh)
74.9
ZGLM-5.3 (Max)
74.8
AnthropicClaude Opus 4.8 (Max)
74.3
AnthropicClaude Opus 4.7 (Max)
73.6
xGrok 4.5 (High)
72.4
MetaMuse Spark 1.2 (Xhigh)
72.2
AlibabaQwen3.8 2.4T A95B
71.9
AnthropicClaude Sonnet 5 (Max)
71.5
ZGLM 5.3 Flash
71.5

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
1#1AnthropicClaude Fable 5.1 (Max, Default Fallback)Anthropic81.6
2#2OGPT-5.6 Sol (Xhigh)OpenAI78.3
3#3AnthropicClaude Opus 5 (Max)Anthropic78
4#4OGPT-6 Astra (max)OpenAI76.9
5#5xGrok 4.6 (High)xAI76.8
6#6OGPT-5.6 Terra (Max)OpenAI76.7
7#7AnthropicClaude Fable 5 (Max, Opus 4.8 Fallback)Anthropic76.5
8#8GoogleGemini 3.8 Flash (High)Google76.3
9#9Moonshot AIKimi K3 (Max)Moonshot AI76.2
10#10AlibabaQwen3.8 Max (0902)Alibaba76.2
11#11GoogleGemini 3.7 Flash (High)Google76.1
12#12OGPT-5.5 (Xhigh)OpenAI74.9
13#13ZGLM-5.3 (Max)Z.ai74.8
14#14AnthropicClaude Opus 4.8 (Max)Anthropic74.3
15#15AnthropicClaude Opus 4.7 (Max)Anthropic73.6
16#16xGrok 4.5 (High)xAI72.4
17#17MetaMuse Spark 1.2 (Xhigh)Meta72.2
18#18AlibabaQwen3.8 2.4T A95BAlibaba71.9
19#19AnthropicClaude Sonnet 5 (Max)Anthropic71.5
20#20ZGLM 5.3 FlashZ.ai71.5
21#21OGPT-5.6 Luna (Max)OpenAI71.4
22#22MetaMuse Spark 1.1 (Xhigh)Meta71.3
23#23OGPT-5.4 (Xhigh)OpenAI71.1
24#24GoogleGemini 3.5 Flash (High)Google70.1
25#25GoogleGemini 3.6 Flash (High)Google69.2
26#26DeepSeekDeepSeek V4 Flash 0731 (Max)DeepSeek69.1
27#27AlibabaQwen3.8 MaxAlibaba68.9
28#28ZGLM-5.2 (Max)Z.ai68.8
29#29GoogleGemini 3.1 Pro PreviewGoogle68.8
30#30DeepSeekDeepSeek V4 Pro 0813 (Max)DeepSeek68.8
31#31AlibabaQwen3.8 27B (Xhigh)Alibaba68.1
32#32AlibabaQwen3.7 MaxAlibaba66
33#33AnthropicClaude Sonnet 4.6 (Max)Anthropic63
34#34Moonshot AIKimi K2.6 (Reasoning)Moonshot AI61.8
35#35Moonshot AIKimi K2.7 CodeMoonshot AI60.8
36#36XiaomiMiMo-V2.5-Pro (Reasoning)Xiaomi60.2
37#37KKAT Coder Pro V2Kwaipilot59.5
38#38DeepSeekDeepSeek V4 Pro 0424 (Max)DeepSeek59.4
39#39NNex-N2-Pro (Based on Qwen3.5-397B-A17B)Nex Agi59.1
40#40THy3Tencent58.8
41#41MiniMaxMiniMax-M3MiniMax58.6
42#42iLing-3.0-flash-VLinclusionAI57
43#43XiaomiMiMo-V2.5Xiaomi56.8
44#44DeepSeekDeepSeek V4 Flash 0420 (Max)DeepSeek56.2
45#45OGPT-5.4 nano (Xhigh)OpenAI56.1
46#46OGPT-5.4 mini (Xhigh)OpenAI56.1
47#47AlibabaQwen3.7 PlusAlibaba55.9
48#48ZGLM-5.1 (Reasoning)Z.ai55.8
49#49iLing-3.0-flash-FininclusionAI55.6
50#50AlibabaQwen3.6 PlusAlibaba54.5
51#51AlibabaQwen3.6 27B (Reasoning)Alibaba53.7
52#52TInkling SmallThinkingmachines52.9
53#53USolar Pro 4Upstage52.7
54#54MiniMaxMiniMax-M2.7MiniMax52.6
55#55AnthropicClaude 4.5 Sonnet (Reasoning)Anthropic52.1
56#56TInkling (Xhigh)Thinkingmachines52.1
57#57xGrok Build 0.1 0616xAI51.5
58#58iLing 3.0 FlashinclusionAI50.6
59#59XiaomiMiMo-V2-Flash (Non-reasoning)Xiaomi49.8
60#60OGPT-5.1 (High)OpenAI49.4

Source: Artificial Analysis (artificialanalysis.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