Filter

×
Active Filters Clear All
Keyword: 预训练 ×
8 Total Reports
Google Other 2026-07-24

Google Begins Gemini 4 Pre-training with 4M+ Context and Monthly Releases

Alphabet confirms start of largest pre-training run for Gemini 4, featuring 4M+ context and native multimodality with near-monthly releases. 2026 capex raised to $195-205B, Google Cloud Q2 up 82%, signaling full-stack AI acceleration.

NVIDIA Other 2026-07-19

NVIDIA Vera Rubin Platform and Dynamo 1.0 Disaggregate Inference, Shift Focus to Intelligence per Dollar

NVIDIA unveils Vera Rubin platform with a 7-chip stack (Vera CPU, Rubin GPU, NVLink 6, etc.) and Dynamo 1.0 inference disaggregation. A single NVL72 rack packs 72 GPUs/36 CPUs with 1.6 PB/s bandwidth, achieving up to 7x inference performance. The new 'intelligence per dollar' metric signals a shift from training to inference cost competition.

Google Other 2026-07-09

Google Gemini 3.5 Pro Rebuilds from Scratch: 2M Token Context Window Reshapes AI Frontier

Google DeepMind targets July 17 for Gemini 3.5 Pro, a full architectural rewrite of its pretraining stack to overcome deficits in math reasoning, SVG generation, and image quality. Specs include a 2M token context window, Deep Think reasoning layer, and multi-step autonomous workflows, though unconfirmed by Google.

Anthropic Other 2026-07-07

Anthropic企业AI采用首超OpenAI 300亿年化收入运行率确认

...

OpenAI Other 2026-07-05

OpenAI Winds Down Fine-Tuning API: A Strategic Shift in AI Customization Landscape

OpenAI plans to phase out its fine-tuning API by 2027, stopping new task creation but allowing inference on existing models. This forces startups relying on fine-tuning for differentiation to migrate to open-source models or RAG, reshaping the AI customization ecosystem.

Meta Other High Signal 2026-03-11

Meta Accelerates Custom AI Chip Roadmap with Focus on Inference Optimization

Meta plans to launch four generations of MTIA AI chips in two years, adopting an 'inference-first' design strategy optimized for generative AI tasks. Built on PyTorch and open standards, the chips enable seamless data center deployment, targeting improved compute efficiency and cost control.

NVIDIA Other High Signal 2026-03-01

NVIDIA Releases Agentic AI Blueprint and Inference Models for Telecom

NVIDIA introduces Agentic AI blueprint and specialized inference models for telecom, built on NeMo framework to autonomously handle network operations. The solution lowers deployment barriers through pre-trained models, advancing telecom networks toward autonomous architecture.

Huawei Other 1970-01-01

Huawei Ascend 910C Trains 1.6T-Parameter MoE Model: First Full Pipeline on Domestic AI Chips

Huawei, in collaboration with research institutes, completed full-parameter post-training of DeepSeek-V4-Pro (1.6 trillion parameters, MoE) on an Ascend 910C cluster. Key metrics: stable 1,500 steps on 1,000 cards, 30% compute utilization, 14% operator efficiency gain, zero reliance on foreign GPUs. This marks the first end-to-end trillion-parameter training loop on domestic chips.