AI × GTM News

Gong introduced a redesigned seller workspace featuring a built-in AI Assistant, new performance dashboards, and Gong Enrich for automated contact data management.

Gong

Mistral AI releases its latest large language model, Mistral Large 24.11 (codenamed Le Chonk), available for deployment and testing on the Hugging Face platform.

Vercel Flags now initializes 7x faster on cold starts and optimizes usage reporting to prevent extended function durations.

Vercel

Mistral AI has released the weights for Mistral Large 4, a 1-trillion parameter multimodal open-weight model optimized for coding and reasoning.

Mistral AI

Vercel has released a plugin for Google Antigravity to help AI agents use Vercel APIs, AI SDK, and Vercel Functions to build and deploy applications.

VercelGoogle

LangChain updates Managed Deep Agents to include browser-based automation for sites without APIs and asynchronous subagents for background tasks.

LangChain

IBM releases Granite Switch 4.1.1, a new iteration of lightweight models designed for efficient software-like updates and enterprise deployment.

Hugging Face has launched open-source models optimized for speech recognition and decision-making specifically designed to run on Apple hardware.

LangChain adds native browser capabilities to its managed agent platform, allowing agents to navigate websites and hand off sessions for manual authentication.

LangChain

Vercel Firewall now allows users to separately log or block AI crawlers for model training and AI assistants for user queries.

Vercel

ElevenLabs has launched SFX v3 in ElevenCreative and its API, featuring improved multi-sound prompts, music, vocals, and audio clarity.

ElevenLabs

OpenAI released an updated Model Spec defining five core values to guide model judgment, agency, and alignment with human intent.

OpenAI

Strix open-source AI penetration testing tool now supports Vercel AI Gateway, enabling access to 400+ models through a single API.

VercelStrix

Vercel now provides city-level data for nearly all requests and postal-code data for 90% of traffic with more frequent data refreshes.

Vercel

Node.js 26 LTS is now generally available for Vercel builds and functions, featuring HTTP/2 by default and the new Temporal API.

Vercel

Vercel Connect now supports webhook triggers for Resend, X, Twilio, Notion, and Telegram to forward events to Vercel projects.

Vercel

Hugging Face introduces Project ProtoStar to standardize and prepare publicly funded scientific data for use in artificial intelligence training and research.

Vercel CI allows developers to run Turborepo CI tasks with one line of configuration, optimizing build times for coding agents by running only affected tasks.

Vercel

Google DeepMind has introduced Gemini 4 Carbon, its latest high-performance frontier AI model.

Virtual Models provide an abstraction layer for model configurations, allowing developers to update models and fallbacks centrally without code changes.

Vercel

Developers can now use custom virtual model IDs to manage model settings and provider fallbacks centrally within the Vercel AI Gateway.

Vercel

Particle-1, a model post-trained on DeepSeek V4.1 Flash for software engineering tasks, is now available for use via Vercel AI Gateway.

VercelParticle AI

Vercel CI is a new tool designed to speed up pull request workflows by running only the specific tasks affected by code changes.

Vercel

NVIDIA released new multilingual transcription models on Hugging Face that support prompting and adaptation for specific transcription tasks.

Box Mount enables agents running in Vercel Sandbox to read and write Box files using standard filesystem operations with automatic synchronization.

VercelBox

OpenAI and ServiceNow are integrating OpenAI models into ServiceNow workflows and bringing ServiceNow capabilities into ChatGPT and Codex.

OpenAIServiceNow

ElevenLabs released Cue, an AI model providing character-level timestamps and rapid audio-to-text alignment for recordings of any length.

ElevenLabs

Sierra released a new tool providing guardrails and controls for AI agents to ensure compliance across various channels and jurisdictions.

Sierra

Paseo has integrated the Vercel Sandbox plugin, allowing users to run and persist agent sessions for fx, Codex, Claude Code, and others in the cloud.

VercelPaseo

ElevenLabs signed a Memorandum of Understanding with the Government of Karnataka to deploy voice AI for citizen services and voice restoration projects.

ElevenLabsGovernment of Karnataka

LangChain introduces a new SDK for Managed Deep Agents to enable proactive, scheduled agentic workflows beyond simple cron triggers.

LangChain

Vercel has increased the compressed image layer limit from 500 MB to 2 GB to support larger toolchains for coding-agent evaluations.

Vercel

ElevenLabs has partnered with the Spanish ALS association adELA to offer free AI-driven voice restoration services to patients living with ALS.

ElevenLabsadELA

Google DeepMind released Genie 4, a new generation world model capable of generating realistic real-time simulated environments.

Clay introduced a post-trained model router that automatically selects the optimal AI model for tasks to reduce costs by 47%.

ClayApplied Compute

The fx adapter now runs harness and model requests in the same process as applications, improving security by keeping credentials on the host.

Vercel

Writer introduced invisible, machine-readable watermarks for its Palmyra models, allowing enterprises to track and identify AI-generated content.

Writer

Writer released new features allowing enterprises to own content watermarks and implement governance controls across AI agents and generated content.

Writer

Users can now track and query how AI bots and crawlers access their sites, including request counts by bot and most-crawled pages for model training.

Vercel

OpenAI and Oracle are deepening their collaboration to bring enterprise applications and AI agents into ChatGPT while powering new Oracle AI offerings.

OpenAIOracle