1
LLM Batch APIs: The Half-Price Lane Nobody Budgets
大模型批量API是省钱利器，非实时任务可享半价，适合高吞吐场景。
Article URL: https://www.digitalapplied.com/blog/llm-batch-api-pricing-landscape-2026 Comments URL: https://news.ycombinator.com/item?id=49299210 Poin…
2
From Compression to Deployment: Real-Time and Energy-Efficient FastGRNN on Ultra-Constrained Microcontrollers
从模型压缩到微控制器部署，突破超低功耗设备实时运行RNN的瓶颈
arXiv:2606.17249v1 Announce Type: cross Abstract: The dominant trajectory of modern machine learning has been to scale up: larger models, larger accel…
3
Real-time LLM Inference on Standard GPUs (3k tokens/s per request)
标准GPU上实现每秒3000 tokens的实时LLM推理，突破速度瓶颈，为AI Agent落地提供硬核方案。
Article URL: https://blog.kog.ai/real-time-llm-inference-on-standard-gpus-3-000-tokens-s-per-request/ Comments URL: https://news.ycombinator.com/item?…
4
Starchild-1 by Odyssey
实时理解多模态世界的AI模型，开创实时世界模型新范式
The first real-time multimodal world model Discussion | Link