← 回總覽

Meta 因性能问题推迟 'Avocado' AI 模型

📅 2026-03-13 09:45 The Rundown AI 人工智能 2 分鐘 1725 字 評分: 82
Meta Avocado 模型 Llama 4 AI 路线图 Watermelon 模型
📌 一句话摘要 据报道,Meta 已推迟其 'Avocado' 模型,该模型性能介于 Llama 4 和顶级前沿模型之间,同时将重心转向代号为 'Watermelon' 的更大项目。 📝 详细摘要 这条推文揭示了 Meta AI 发展路线图上的一项重要更新。根据引用的独家爆料,Meta 因性能问题正在推迟其 'Avocado' 模型,因为它目前的性能介于 Llama 4 和即将推出的 Gemini 2.5/3 等前沿模型之间存在差距。报告还提到 'Watermelon' 是一个更具雄心的继任者。主推文强调了围绕此次泄密事件在内部或社区中引发的高度讨论。 📊 文章信息 AI 评分:82 来
![Image 1: The Rundown AI](https://www.bestblogs.dev/en/tweets?sourceId=SOURCE_045a9f)

This must be blowing up on Moltbook right now

!Image 2: Eli Tan

#### Eli Tan

@elitanjourno · 11h ago

Scoop from me: Meta delays Avocado 🥑 after performance concerns. The new model is better than Llama 4 but worse than other frontiers — somewhere in between Gemini 2.5 and 3.

There is excitement about the model after Avo, named after an even larger fruit: Watermelon 🍉

!Image 3: Tweet image

7

8

84

51.8K

5 Replies

1 Retweets

17 Likes

4,203 Views ![Image 4: The Rundown AI](https://www.bestblogs.dev/en/tweets?sourceid=045a9f)

One Sentence Summary

Meta has reportedly delayed its 'Avocado' model, which ranks between Llama 4 and top-tier frontier models, while shifting focus to a larger project codenamed 'Watermelon'.

Summary

This tweet highlights a significant update in Meta's AI development roadmap. According to the quoted scoop, Meta is delaying its 'Avocado' model due to performance concerns, as it currently sits in a performance gap between Llama 4 and upcoming frontier models like Gemini 2.5/3. The report also mentions 'Watermelon' as a more ambitious successor. The main tweet emphasizes the high level of internal or community discussion surrounding this leak.

AI Score

82

Influence Score 5

Published At Today

Language

English

Tags

Meta

Avocado Model

Llama 4

AI Roadmap

Watermelon Model

查看原文 → 發佈: 2026-03-13 09:45:11 收錄: 2026-03-13 12:00:35

🤖 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。