Ειδήσεις: OPEN_WEIGHT | PXAI

07/04 05:43 dev.to

Gemma 4 MoE: frontier quality at 1/10th the API cost

Gemma 4 MoE LLM open weights cost savings inference

02/04 19:00 engadget.com

Google releases Gemma 4, a family of open models built off of Gemini 3

Google Gemma 4 open-weight models Gemini 3 AI technology

30/03 20:01 venturebeat.com

Cohere's open-weight ASR model hits 5.4% word error rate — low enough to replace speech APIs in production pipelines

Cohere ASR open-weight model low WER speech API data residency

26/03 23:32 dev.to

Local LLM Acceleration: Quantization, TTS, and 1M Tokens/Sec

Mistral AI Voxtral TTS open weights quantization 1M tokens/s local LLM

26/03 21:41 x.com

TechPulseBytes: [View all]( https:// github.blog/latest/) Open weights change the game. Researchers can now audit, fine-tune, and deploy...

open weights AI transparency model audit fine‑tuning API independence competitive ecosystem

25/03 02:57 dev.to

Ai2 Releases MolmoWeb: A Game-Changer for Visual Web Agents

MolmoWeb AI agents web navigation open-weight framework visual web agents Ai2

24/03 17:00 venturebeat.com

Ai2 releases MolmoWeb, an open-weight visual web agent with 30K human task trajectories and a full training stack

Ai2 MolmoWeb open-weight visual web agent MolmoWebMix human task trajectories

12/03 01:42 venturebeat.com

Nvidia's new open weights Nemotron 3 super combines three different architectures to beat gpt-oss and Qwen in throughput

Nvidia Nemotron 3 Super hybrid architecture state‑space transformers latent mixture‑of‑experts

12/03 01:42 venturebeat.com

Nvidia's new open weights Nemotron 3 super combines three different architectures to beat gpt-oss and Qwen in throughput

Nvidia Nemotron 3 Super hybrid architecture state‑space transformers latent mixture‑of‑experts

Loading...