Baidu Releases ERNIE 5.5 Flash Preview with 280B Total Parameters and 14B Activated per Token
Baidu has released ERNIE 5.5 Flash Preview, with 280B total parameters and only 14B activated per token. Arena.ai says the model, at 1668 points on the Code Arena WebDev leaderboard, is trending toward becoming the second-ranked open-source model, and its weights will be open-sourced under Apache 2.0. Its score is slightly above DeepSeek-V4.1-Flash (Max) at 1664 points and Kimi K3 (Max) at 1658 points, and on par with Qwen3.8-Max at 1670 points.
Why read this
SelectedAI score 70
Baidu has disclosed the parameter count and activation scale of ERNIE 5.5 Flash Preview, along with its position on Code Arena, making it easy to compare against open-source models in the same tier.
Topics
View original postErnie 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!
View original
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!
Body
The latest from @arena 🚀
ERNIE 5.5 Flash Preview — 280B total parameters, just 14B active per token. See how it stacks up.
Thanks to the community for putting it to the test! 🙌