AITOPAI热报
ModelsOfficial

Baidu 发布 ERNIE 5.5 Flash Base：280B 参数 MoE 基座模型

Baidu Releases ERNIE 5.5 Flash Base: A 280B MoE Base Model With 14B Active Parameters

官方一手marktechpost@Michal Sutter
Blog/Media
Read the original
Why we picked it

百度把 280B 的 ERNIE 5.5 Flash 基座模型在还没做后训练前就放出来了，想自己微调、做后训练研究的开发者可以拿来玩。

Baidu 发布了 ERNIE 5.5 Flash Base，这是一个总参数 280B 的稀疏 MoE 基座模型，每次推理仅激活 14B 参数。该模型在 33.5T tokens 上完成预训练，且在任何中期训练或后训练（如指令微调、RLHF）之前就直接开源发布。这意味着开发者可以基于原始基座自行开展微调和后训练。

Image · marktechpost
Original · marktechpost

Baidu Releases ERNIE 5.5 Flash Base: A 280B MoE Base Model With 14B Active Parameters

Baidu has released ERNIE 5.5 Flash Base, a 280B sparse MoE base model with 14B active parameters, trained on 33.5T tokens and published before any mid- or post-training. The post Baidu Releases ERNIE 5.5 Flash Base: A 280B MoE Base Model With 14B Active Parameters appeared first on MarkTechPost .

#ERNIE#Baidu#MoE#开源模型#基座模型

Keep reading

Recent stories with the same tag or section
ModelsUS vs ChinaMulti-sourceSelected00:37
vLLM ships day-0 support for Baidu's ERNIE 5.5 Flash Base

The benchmark gap over DeepSeek-V4-Flash-Base in the same evaluation setup makes this one of the stronger open base weights released for coding workloads, and day-0 vLLM support means teams can deploy it without custom serving work. Practitioners running RL research or domain adaptation get open base weights plus serving recipes for expert parallelism, KV offload and prefill/decode disaggregation on GB200 and H100 hardware.

#ERNIE 5.5 Flash#Baidu#vLLM#open-source models
Event
官方账号vLLM@vllm_project11 sourcesSource
ModelsUS vs ChinaSelected00:29
Baidu releases ERNIE 5.5 Flash Base with open weights

Releasing a pure pre-training checkpoint lets researchers and engineers run their own mid- and post-training experiments from the same starting point, making method comparisons reproducible. The 1M-token context and FP8 weights make it practical for long-document fine-tuning on a single cluster. Teams benchmarking open base models against Qwen or DeepSeek checkpoints now have a new Baidu entry to include.

#ERNIE#Baidu#open-source models#MoE
官方账号百度 AI Baidu@Baidu_IncSource
ModelsBuildersSelected21:23
Baidu open-sources ERNIE 5.5 Flash Base for continued training

The base-model-only release signals Baidu competing for the fine-tuning and localisation market rather than end users, letting teams inject industry knowledge or adapt regional languages on top of a large pretrained backbone. Developers evaluating against Qwen or DeepSeek open weights get a sparse 5%-activation design that keeps serving costs low at 1M-token contexts. Availability details and licence terms for ASEAN teams remain worth checking before adoption.

#ERNIE#Baidu#open-source models#fine-tuning
小互@imxiaohuSource