AI
JetBrains Opens Mellum2 to Cut Coding Agent Call Costs
JetBrains Mellum2 is a 12B MoE model that activates 2.5B parameters per token. It adds a private coding model option for IDE and agent workflows.
AI
NVIDIA Bundles a 550B Open Model With an Agent Runtime Stack
NVIDIA introduced Nemotron 3 Ultra alongside NemoClaw, OpenShell, and CUDA-X agent skills, pushing open agent competition into the runtime layer.
AI
OpenAI models and Codex are now generally available on Amazon Bedrock
OpenAI GPT-5.5, GPT-5.4, and Codex are generally available on Amazon Bedrock, shifting authentication, billing, and feature boundaries into AWS.
AI
NVIDIA opens a 32B robotaxi model and closed-loop training stack
NVIDIA Alpamayo 2 Super ties a 32B VLA teacher model to AlpaGym, OmniDreams, CoC auto-labeling, and agent skills for L4 robotaxi development.
AI
QVAC 0.12.0 brings TurboQuant to local KV cache pressure
Tether announced QVAC SDK 0.12.0 with TurboQuant support. The useful question is how KV cache compression changes local long-context AI.
AI
AgentCore A/B testing turns prompt edits into release experiments
AWS AgentCore Optimization preview productizes agent quality loops with trace-based recommendations, batch evals, and A/B testing.
AI
Cohere Command A+ ships as an open-weight MoE for two H100s
Cohere Command A+ combines Apache 2.0 open weights, a 218B MoE design, 25B active parameters, 128K context, and enterprise deployment options.
AI
SageMaker Adds OpenAI API Support for AWS-Hosted Models
AWS SageMaker now supports /openai/v1 endpoints, lowering the migration cost for OpenAI SDK, LangChain, Strands Agents, and AI gateways.
AI
MiniMax M3 brings 1M context to open-weight coding models
MiniMax M3 combines 1M context, multimodality, and coding-agent benchmarks, but its weights and technical report are still pending verification.
AI
Mistral Search Toolkit separates RAG failures from model failures
Mistral Search Toolkit public preview treats RAG and agent-search failures as retrieval, pipeline, and evaluation problems.
AI
RTX Spark Debuts as a Local AI PC for 120B LLMs
NVIDIA and Microsoft introduced RTX Spark, a Windows PC category for local agents with 120B LLMs, 128GB unified memory, and OpenShell.
AI
CoreWeave Agentic AI Turns Inference Logs Into Training Signals
CoreWeave introduced agentic AI integrations that connect inference, W&B Weave observability, serverless RL, and coding-agent tooling into one improvement loop.