Product Launches — Page 2

NVIDIA released Alpamayo 2 Super, a commercial open model for autonomous driving, under the Linux Foundation's OpenMDW-1.1 license. It ranks first on LingoQA, outperforming GPT-4o, Gemini 2.5 Pro, and Qwen2.5-VL 72B.

Alibaba's Qwen3.8-Max, a 2.4-trillion-parameter MoE model, launched on August 3, 2026, outperforming Claude Fable 5 on coding benchmarks. Open weights are expected the week of August 10, 2026, with competitive API pricing at $2/$6 per million input/output tokens.

The Friend AI Pendant 2.0 relaunches with a speaker, a $249 price, and a $10 monthly subscription for extended conversation memory. Learn about the changes and what they mean for the wearable AI market.

Google DeepMind released Gemini Robotics ER 2, an embodied reasoning model for robotics, on July 30, 2026. It enables multi-step task planning, real-time progress tracking, and multi-robot collaboration, and is publicly available to developers via the Gemini API and Google AI Studio.

Perplexity launches Personal Computer for Windows, extending its AI agent capabilities to the Windows OS. This enterprise-focused solution runs locally, accesses user files, and costs $200/month.

OpenAI launched ChatGPT Health to all US users, integrating medical records and claiming "clinician-level reasoning." This release follows a lawsuit alleging dangerous medical recommendations from ChatGPT.

Lingverse (formerly Ling AI) raised $29 million to build the iKairos, an AI wearable that captures snapshots and uses AI to generate "memory" images, reviving the Jibo robot's legacy.

Halliday's G2 smart glasses, launching in September 2026 for $599, offer AI-powered meeting intelligence without a camera, targeting secure professional environments. This second-generation device aims to compete directly with Even Realities G2.

Anthropic's Claude voice mode now supports Opus and Sonnet models, expanding its capabilities beyond Haiku. The update also integrates voice mode with apps like Gmail, Slack, and Canva, allowing hands-free interaction for complex business tasks and supporting ten languages.

Meta launched Content Seal in July 2026, an invisible watermarking tool for its Muse AI images, diverging from Google's SynthID. This decision impacts AI content detection and interoperability.

Google DeepMind launched Gemini 3.5 Flash Cyber on July 21, 2026, a new AI model that found 55 unique confirmed vulnerabilities in Chrome's V8 JavaScript Engine. This lightweight cybersecurity model is available to governments and trusted partners through Google's CodeMender security agent.

Capital One released VulnHunter, an open-source AI security tool, on July 17, 2026. It uses agentic workflows and an "attacker-first forward analysis" to find exploitable vulnerabilities in source code, enhancing software security.

Google rebrands NotebookLM to Gemini Notebook, integrating it deeper into the Google ecosystem. The update includes a secure cloud computer feature for native code execution, enhancing data analysis capabilities for users.

1Password introduces a zero-exposure credential integration for Anthropic Claude, enabling AI agents to act on your behalf across the web without ever seeing your passwords. This feature is now available on Mac for business, family, and individual plan users.

OpenAI launched the Codex Micro, a $230 programmable macro pad co-designed with Work Louder, on July 15, 2026. This device is designed to manage AI coding agents and marks OpenAI's entry into the hardware market.

PrismML launched Bonsai 27B, the first 27B-class AI model to run on a smartphone, offering multimodal, 262K-token context capabilities. The 1-bit variant is 3.9 GB and runs on an iPhone 17 Pro, while the 5.9 GB ternary variant runs on laptops, both under an Apache 2.0 license.

Apple's new on-device SpeechAnalyzer API achieved a 2.12% word error rate, outperforming OpenAI Whisper Small's 3.74% and running three times faster in an independent benchmark for English transcription.

OpenAI is reportedly developing a screenless, movable ChatGPT smart speaker as its first hardware product, designed to be an AI companion. Slated for a 2027 launch, this device will compete with offerings from Apple, Amazon, and Google.

Mistral AI's new 8-billion-parameter Robostral Navigate model allows robots to navigate complex environments using only a single RGB camera, reducing hardware costs and simplifying autonomous robot deployment.

SpaceXAI (xAI) launched Grok 4.5 on July 8, 2026, an "Opus-class model" with aggressive pricing of $2 per million input tokens and $6 per million output tokens, strategically preceding OpenAI's GPT-5.6 launch.