← 回總覽

OpenAI GPT-5.4 首次亮相 LMSYS 排行榜

📅 2026-03-12 04:14 Arena.ai 人工智能 3 分鐘 3485 字 評分: 88
GPT-5.4 OpenAI LMSYS LLM 基准测试 Document Arena
📌 一句话摘要 GPT-5.4 在 LMSYS 平台的 Document Arena(并列第 2)和 Arena Expert(前 5)中均取得了顶尖排名。 📝 详细摘要 本推文报告了 OpenAI 新发布的 GPT-5.4 模型的初步基准测试结果。其亮点在于强大的文档分析能力,目前与 Claude 3.6 Sonnet 并列第 2 位。此外,它在 “Arena Expert” 分类中位列前 5,并在数学、商业和指令遵循(Instruction Following)等专业领域表现出极强的竞争力。这为 GPT-5.4 的实际能力提供了首个独立的实证证据。 📊 文章信息 AI 评分:88 来
Skip to main content ![Image 1: LogoBestBlogs](https://www.bestblogs.dev/ "BestBlogs.dev")Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters

⌘K

Change language Switch ThemeSign In

Narrow Mode

OpenAI GPT-5.4 Debuts on LMSYS Leaderboards ===========================================

OpenAI GPT-5.4 Debuts on LMSYS Leaderboards =========================================== ![Image 2: Arena.ai](https://www.bestblogs.dev/en/tweets?sourceId=SOURCE_39a65f) ### Arena.ai

@arena

GPT-5.4 by @OpenAI lands tied #2 on Document Arena and in top 5 for Arena Expert.

Document Arena Highlight:

  • #2 tied with Claude Sonnet 4.6
Text Arena Highlights:
  • #5 for Arena Expert
  • top 10 in Business, Management, & Financial Ops and Writing, Literature, & Language fields
  • top 15 in Math, Instruction Following, Multi-Turn & Hard Prompts
  • top 15 in Text Arena overall
!Image 3: Tweet image

!Image 4: OpenAI

#### OpenAI

@OpenAI · 6d ago

GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT.

GPT-5.4 is also now available in the API and Codex.

GPT-5.4 brings our advances in reasoning, coding, and agentic workflows into one frontier model.Show More

!Image 5: Tweet image

1,690

3,295

23.5K

6.5M

Mar 11, 2026, 8:14 PM View on X

2 Replies

5 Retweets

78 Likes

6,460 Views ![Image 6: Arena.ai](https://www.bestblogs.dev/en/tweets?sourceid=39a65f) Arena.ai @arena

One Sentence Summary

GPT-5.4 achieves top rankings in Document Arena (#2) and Arena Expert (top 5) on the LMSYS platform.

Summary

This tweet reports the initial benchmarking results for OpenAI's newly released GPT-5.4 model (as referenced in the quoted tweet). It highlights strong performance in document analysis, where it is tied for 2nd place with Claude 3.6 Sonnet. Furthermore, it ranks in the top 5 for 'Arena Expert' and shows competitive performance across specialized fields including Math, Business, and Instruction Following. This provides the first independent empirical evidence of GPT-5.4's capabilities.

AI Score

88

Influence Score 9

Published At Today

Language

English

Tags

GPT-5.4

OpenAI

LMSYS

LLM Benchmarking

Document Arena HomeArticlesPodcastsVideosTweets

OpenAI GPT-5.4 Debuts on LMSYS Leaderboards | BestBlogs.dev ===============

查看原文 → 發佈: 2026-03-12 04:14:55 收錄: 2026-03-12 06:00:56

🤖 問 AI

針對這篇文章提問,AI 會根據文章內容回答。按 Ctrl+Enter 送出。