百度 Ernie 5.5 Flash Preview 逼近 Code Arena 开源模型第二
百度这次发了 ERNIE 5.5 Flash Preview，280B 参数只激活 14B，1M 上下文，速度超 200 tokens/sec。它的 WebDev 分数已经追平 Qwen3.8-Max，压过 DeepSeek-V4.1-Flash。权重 11 月 25 日前用 Apache 2.0 开源，想本地部署写代码的可以盯一下。
百度发布 ERNIE 5.5 Flash Preview，在 Code Arena: WebDev 排行榜拿到 1668 分，正接近开源模型第二名，超过 DeepSeek-V4.1-Flash (Max) 的 1664 分和 Kimi K3 (Max) 的 1658 分，与 Qwen3.8-Max 的 1670 分基本持平。模型权重计划于 11 月 25 日前以 Apache 2.0 协议开源。
该模型总参数量 280B，每 token 激活 14B，采用为高效推理设计的新架构，解码速度超过 200 tokens/sec，支持 1M token 上下文窗口。官方称其在 Terminal-Bench、AutomationBench 等基准上表现有竞争力，能力覆盖编程、3D 建模、知识工作和内容创作。
凭 1668 分，ERNIE 5.5 Flash Preview 与 Qwen、DeepSeek、Kimi 系列的开源旗舰在 WebDev 场景直接竞争，且开源协议宽松，对需要本地部署的开发者来说多了一个 Apache 2.0 权重的选项。
Ernie 5.5 Flash Preview by @Baidu_Inc is trending towards becoming the #2 open model on the Code Arena: WebDev, at 1668 pts. The weights will be open-sourced under Apache 2.0 soon. Its score puts it just above: - DeepSeek-V4.1-Flash (Max) at 1664 - Kimi K3 (Max) at 1658 and on par with: - Qwen3.8-Max at 1670 Congrats to the @Baidu_Inc team on this release! Baidu Inc. @Baidu_Inc ERNIE 5.5 Flash Preview just landed! 🔥 Built on a new architecture designed for efficient inference, the latest open-weight foundation model in the ERNIE family delivers stronger performance in coding, 3D modeling, knowledge work and content creation. It keeps decoding fast while posting competitive scores on Terminal-Bench, AutomationBench, and other benchmarks. Here's the rundown: - 280B total parameters, 14B active per token - 1M-token context window - Generation speed above 200 tokens/sec Model weights will be released by November 25. 🔗 View Quoted Tweet📊 10 ⚡
This report is categorized under AITOP's "Models" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include ERNIE, Baidu, DeepSeek, Qwen.
Source attribution: curated from lmarena.ai (@lmarena_ai).
