Baidu open-sources ERNIE 5.5 Flash Base, a 280B-parameter MoE foundation model shipped without post-training
Baidu_Inc · x · 2026-11-06
Baidu released ERNIE-5.5-Flash-Base, an open foundation model deliberately shipped without fine-tuning or post-training, leaving developers to decide its direction; a post-trained version is planned.
- MoE architecture with 280B total parameters, 14B active per input; pre-trained on 33.5T tokens with 128K training context, designed to support up to 1M context at fine-tuning and inference
- Baidu says it competes with similarly sized open base models like DeepSeek V4 Flash and MiMo V2.5 on reasoning, code, and long-context benchmarks
- Weights available on Hugging Face in FP8 and BF16
More from Models
- Miles Brundage: Inkling 2 Small is good — Miles_Brundage · 2026-11-06
- Users claim Claude performs better with an empty system prompt than Anthropic's official one — repligate · 2026-11-06
- Palisade's Jeff Ladish: models better at everything but still not funny or persuasive — JeffLadish · 2026-11-06
- Palisade's Jeff Ladish calls to revive METR time-horizon estimates for new models — JeffLadish · 2026-11-06
- Leaker DM claims Qwen4 is dropping, but even the timeline is a guessing game — bclavie · 2026-11-06
- Artificial Analysis launches Endpoint Accuracy Index v1.1: nearly 1 in 4 serverless endpoints score below reference — ArtificialAnlys · 2026-11-06