Understand this story
On November 5, 2026, Baidu officially open-sourced the base model ERNIE 5.5 Flash Base. According to a same-day report by Leiphone, the model adopts a MoE architecture with a total of 280 billion parameters, activating 14 billion parameters per token, and was pre-trained on 33.5 trillion tokens; it achieves SOTA among models of the same size on benchmarks covering knowledge, reasoning, mathematics, code, multilingual capability, and long context, surpassing models such as DeepSeek-V4-Flash-Base, Hy3-preview-Base, and MiMo-V2.5-Base; its pre-training context length reaches up to 128K, and the fine-tuning and inference stages support a context of up to 1 million tokens. Baidu officially posted a tweet on November 6 confirming that ERNIE 5.5 Flash Base has a parameter scale of 280 billion and is provided as an open-source base model, along with a release link. The model has currently been open-sourced and released, with no further developments at this time.
AI-generated from public reports
Latest development · Report
Baidu Open-Sources the 280-Billion-Parameter ERNIE 5.5 Flash Base Foundation ModelDevelopments
1 confirmed developments
- Source:English text is being prepared.
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 6
BaiduFirst-partySelected
Baidu Open-Sources the 280-Billion-Parameter ERNIE 5.5 Flash Base Foundation Model
Baidu released ERNIE 5.5 Flash Base, with a parameter scale of 280 billion, provided as an open-source foundation model. The official tweet also included a release link.
MarkTechPostRelated discussion
MarkTechPost joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
The Decoder
Baidu Open-Sources 280-Billion-Parameter ERNIE 5.5 Flash Base Foundation Model
Baidu has released the open-source foundation model ERNIE-5.5-Flash-Base, deliberately without fine-tuning or post-training, leaving developers to decide the subsequent direction themselves, with post-trained versions to be released later. The model uses a MoE architecture, with 280 billion total parameters and 14 billion activated per input, pre-trained on 33.5 trillion tokens, with a training context of up to 128,000 tokens, and is designed to support fine-tuning and inference-time context of up to 1 million tokens. Baidu says it competes on reasoning, code, and long-context benchmarks with same-scale open-source foundation models such as DeepSeek V4 Flash and MiMo V2.5, and the weights are already available on Hugging Face in FP8 and BF16.
Nov 5
Yicai
AI Evolution Express: Baidu Open-Sources ERNIE 5.5 Flash Base, Kimi Releases Hosted Agents
Baidu has officially open-sourced its latest foundation model, ERNIE 5.5 Flash Base, and Moonshot AI has released hosted agents, Kimi Hosted Agents. Qwen AI Office has upgraded its in-app editing and template capabilities, allowing Word, Excel, and PPT files to be modified as needed; Kling AI has announced the full rollout of Kling 4.0 Flash, deeply integrated into professional film and television workflows. OpenAI says more than half of its revenue comes from enterprise customers and will be ready for an IPO in 2027.
Shao MengRelated discussion
Shao Meng joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
IT Home · AI
Baidu Wenxin open-sources the ERNIE 5.5 Flash Base foundation model, with 280B total parameters
Baidu Wenxin has released and open-sourced ERNIE-5.5-Flash-Base, a sparse MoE language model with 280B total parameters and 14B activated per token, pretrained from scratch on 33.5T tokens. These weights have not undergone mid-training or post-training and can serve as a starting point for domain adaptation, instruction fine-tuning, and reinforcement learning research; fine-tuning and inference support a 1M-token context. The official statement says that the post-trained ERNIE 5.5 Flash will be released subsequently, and it is compared against DeepSeek-V4-Flash-Base, Hy3-preview-Base, and MiMo-V2.5-Base under the same framework across 30 benchmarks.
IT Home · AIRelated discussion
IT Home · AI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Shao Meng
Baidu Open-Sources ERNIE 5.5 Flash Base, a 280B-Parameter MoE Architecture
Baidu has open-sourced and released ERNIE 5.5 Flash Base, a MoE architecture model with 280B total parameters and 14B activated per token. Baidu Research claims it has reached SOTA at the same size and teases that the full version of ERNIE 5.5 Flash is coming soon. The reposter points out that its benchmark compares models of the same parameter scale rather than each company's latest versions, and lists DeepSeek-V4-Flash-Base corresponding to V4.1-Flash-1017, Hy3-preview-Base corresponding to Hy4, and MiMo-V2.5-Base corresponding to MiMo-V2.6-Flash for reference.
Adina YakupRelated discussion
Adina Yakup joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
QbitAI
Baidu Open-Sources ERNIE 5.5 Flash Base, Setting SOTA for Same-Size Models on Multiple Benchmarks
On November 5, Baidu officially open-sourced the base model ERNIE 5.5 Flash Base, which adopts a MoE architecture with a total parameter count of 280 billion and 14 billion parameters activated per token. The model was pre-trained on 33.5 trillion tokens, with a maximum pre-training context length of 128K, and supports a context length of up to 1 million tokens during fine-tuning and inference. The company claims it achieves SOTA among same-size models on benchmarks covering knowledge, reasoning, mathematics, code, multilingual capabilities, and long context, surpassing models such as DeepSeek-V4-Flash-Base, Hy3-preview-Base, and MiMo-V2.5-Base.
LeiphoneSelected
Baidu Open-Sources ERNIE 5.5 Flash Base, Setting SOTA for Same-Size Models Across Multiple Benchmarks
Baidu has officially open-sourced the base model ERNIE 5.5 Flash Base, which adopts a MoE architecture with 280 billion total parameters, activating 14 billion parameters per token, and was pretrained on 33.5 trillion tokens. The model achieves SOTA among same-size models on benchmarks covering knowledge, reasoning, mathematics, code, multilingual capability, and long context, surpassing models such as DeepSeek-V4-Flash-Base, Hy3-preview-Base, and MiMo-V2.5-Base. The pretraining context length reaches up to 128K, and the fine-tuning and inference stages support a context of up to 1 million tokens.
Hacker News Trending
Baidu Releases ERNIE-5.5-Flash-Base Foundation Model with 280B Total Parameters and 14B Activated
The Baidu ERNIE team has released ERNIE-5.5-Flash-Base, a sparse MoE language model with 280B total parameters and 14B activated per token, pretrained from scratch on 33.5T tokens, with a maximum training context of 128K and support for 1M-token context in fine-tuning and inference. The released weights are the checkpoint at the end of pretraining, without mid-training or post-training, and are available on Hugging Face in FP8 and BF16 safetensors formats for domain adaptation, instruction fine-tuning, and reinforcement learning research. The official comparison with DeepSeek-V4-Flash-Base, Hy3-preview-Base, and MiMo-V2.5-Base uses the same evaluation framework: MMLU 93.3, MMLU-Pro 78.4, HumanEval+ 82.3, and LiveCodeBench 40.1.
Heat trend for this story
- Current
- 57
- Period peak
- 61 Nov 6 11:00
- 24-hour change
- —
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.