Agents
-
Agents
OpenAI gives ChatGPT and Codex a structured WebMCP path into websites
OpenAI adds WebMCP-based Site tools to ChatGPT and Codex, letting browser agents invoke structured website actions inside live authenticated sessions.
-
Research
AdaptRubric gives GUI agents task-specific rules for judging success
AdaptRubric builds task-adaptive rubrics for GUI-agent evaluation, with reported gains in reward-model F1 and reinforcement-learning task success.
-
New Products
Perplexity moves its Computer agent onto local NVIDIA hardware
Perplexity Portable Computer runs agent models, orchestration, local search and sandboxed tools on NVIDIA DGX Spark with optional cloud escalation.
-
Research
Recuris improves long-horizon agents by evolving memory, not model weights
Recuris keeps the base LLM frozen while evolving working and experiential memory, with reported gains across long-horizon agent benchmarks.
-
Agents
Google packages Gemini agents for legal and financial workflows
Google Cloud launches Gemini Enterprise for Legal and Financial Services with MCP connectors, A2A APIs, managed agents and governed enterprise controls.
-
Open Source
GPT Image 2 prompt system passes 11K GitHub stars as Agent Skill gains traction
awesome-gpt-image-2 has evolved into a Prompt-as-Code system and Agent Skill, showing growing developer demand for reusable image prompt infrastructure.
-
Open Source
FreeToken runs frontier-scale MoE models on consumer GPUs
FreeToken uses bandwidth-adaptive CPU-GPU execution and semantic caching to serve 35B to 753B MoE models on consumer and workstation hardware.
-
Infrastructure & Chips
NVIDIA puts Groq 3 LPX into production for faster agent inference
NVIDIA has moved Groq 3 LPX into full production with Vera Rubin, targeting decode latency in token-heavy agent workloads and naming Nebius as the first cloud adopter.
-
Research
Inherent trains 27B Faraday agent to replicate research
Inherent’s 27B Faraday agent uses Replica and coding agents to reproduce scientific results, while benchmark caveats complicate direct model comparisons.
-
New Models
DeepSeek adds vision to V4 Flash for multimodal agents
DeepSeek's V4 Flash Vision Exp adds image understanding and reusable file input to its API, expanding low-cost multimodal agent options while remaining experimental.