AI MODELS
NVIDIA Ising Calibration 1.5 Pushes Quantum Computing Forward
NVIDIA's Ising Calibration 1.5 automates quantum processor tuning, outperforming open models. A milestone in AI-driven quantum advancements.
Kimi K3 Beats GPT-5.6 Sol in Cost Efficiency and Coverage
Kimi K3 outperforms GPT-5.6 Sol on cost and multi-attempt coding success, with implications for AI-driven development workflows.
Anthropic Analyzes Claude's Value Variations Across Models
Anthropic's study on Claude reveals how its values shift across models and languages, offering insights into AI behavior and training implications.
OpenAI Launches GPT-5.6: Sol, Terra, Luna Models Go Live
OpenAI's GPT-5.6, featuring Sol, Terra, and Luna models, launches globally with enhanced performance and tiered pricing. Key upgrade for AI workflows.
Anthropic Launches Claude Sonnet 5: Mid-Tier AI with Opus-Level Features
Anthropic's Claude Sonnet 5 debuts as a cost-efficient AI model for coding and workflows, narrowing the gap with higher-end Opus 4.8 models.
Google Rolls Out Nano Banana 2 Lite, Gemini Omni Flash
Google's Nano Banana 2 Lite and Gemini Omni Flash are live, enabling faster, cost-efficient AI image and video generation for developers.
Grok Models Now Available on Databricks Agent Bricks
xAI's Grok models go live on Databricks Agent Bricks, expanding options for enterprises to deploy AI tools alongside their data.
Grok 4.3 Joins Amazon Bedrock, Setting New AI Benchmarks
Grok 4.3, now live on Amazon Bedrock, offers 1M-token context, low hallucination rates, and configurable reasoning for enterprise AI.
Kimi K2.7 Code Outshines Claude Fable 5 at 94% Lower Cost
Kimi K2.7 Code delivers near-parity with Claude Fable 5 for 94% less cost. Open-source proves competitive in generative coding workflows.
US Halts Foreign Access to Anthropic's Fable 5 AI Model, Citing Security
The US government has suspended access to Anthropic's Fable 5 and Mythos 5 AI models for foreign nationals, raising questions about AI export controls.
FSDP and PyTorch Enable Large-Scale Model Training
Fully Sharded Data Parallel (FSDP) in PyTorch, integrated with Ray, optimizes GPU memory usage for scalable training of models like Qwen3-TTS with 1.7B parameters.
MiniMax M3 Debuts on NVIDIA: 1M Token Context, Multimodal AI
MiniMax M3, the 428B-parameter model, launches on NVIDIA infrastructure, offering long-context reasoning and multimodal workflows for enterprise AI.
Anthropic Launches Claude Fable 5, a Safer Mythos-Class AI
Anthropic debuts Claude Fable 5, its first public Mythos-class AI model, combining state-of-the-art performance with new safety safeguards.
NVIDIA's Nemotron 3 Ultra Redefines AI for Long-Running Agents
NVIDIA's Nemotron 3 Ultra, a 550B-parameter AI model, offers faster, cost-efficient reasoning for complex workflows, enabling long-running agents to perform better.
NVIDIA Alpamayo Enhances AV Model Training With Closed-Loop Tech
NVIDIA's Alpamayo introduces closed-loop post-training for autonomous vehicles, advancing reasoning capabilities and deployment readiness.