OpenAI Launches GPT-Live Voice Model That Delegates Complex Tasks to GPT-5.5
OpenAI has replaced ChatGPT's voice mode with GPT-Live, a new voice model that can delegate complex tasks to GPT-5.5 in the background. The previous voice mode was based on a GPT-4o era model with a 2024 knowledge cutoff.
OpenAI Launches GPT-Live Voice Model That Delegates Complex Tasks to GPT-5.5
OpenAI has released GPT-Live, replacing the voice mode in ChatGPT with a new model that automatically delegates complex tasks to GPT-5.5 while maintaining conversation flow.
According to OpenAI, GPT-Live can spin off harder tasks requiring web search, deeper reasoning, or complex work to GPT-5.5 behind the scenes, then bring results back into the conversation. While GPT-5.5 processes these tasks, GPT-Live continues talking to maintain conversation flow.
What Changed
The previous ChatGPT voice mode ran on a GPT-4o era model with a knowledge cutoff sometime in 2024. GPT-Live represents a complete replacement of this system.
OpenAI states it will continuously update the frontier model used by GPT-Live as new models are released. At launch, GPT-5.5 serves as the background model for complex tasks.
Known Issues
During the preview period, the model exhibited a bug where it interrupted users to laugh at statements not intended as jokes. Developer Simon Willison, who had preview access for several weeks in the iPhone app, reported the issue to OpenAI. The company appears to have made adjustments to reduce the frequency of this behavior.
Pricing and technical specifications for GPT-Live have not been disclosed. The model is available in the ChatGPT iPhone app.
What This Means
This release represents OpenAI's approach to the voice assistant problem: using a specialized voice model for real-time conversation while offloading complex reasoning to more capable models. The architecture mirrors how humans might handle conversations—maintaining engagement while mentally processing difficult questions.
The delegation approach addresses a key limitation of previous voice modes: users had to choose between conversational flow and model capability. If the system works as described, it eliminates that tradeoff. However, the lack of disclosed latency metrics for the delegation handoff leaves open questions about how seamless the experience actually is during complex tasks.
Related Articles
OpenAI restores chat sidebar in Mac app after user backlash over confusing redesign
OpenAI has updated its ChatGPT Mac app to restore direct access to chat conversations through a prominent sidebar toggle. The fix addresses user complaints following a July 10 redesign that replaced the native Mac client with an Electron-based app and buried the standard chat interface behind Work and Codex features.
OpenAI's GPT-5.6 Sol Adds Five Reasoning Effort Settings, Follows DeepSeep-R1 RLVR Training Method
OpenAI released GPT-5.6 Sol, a new reasoning model family that comes in three sizes with roughly five to six reasoning-effort settings each. The release follows the DeepSeek-R1 methodology of using reinforcement learning with verifiable rewards (RLVR), nearly two years after OpenAI's original o1 model popularized LLM-based reasoning.
Moonshot AI releases Kimi K3, China's largest model at 2.8 trillion parameters
Beijing-based Moonshot AI released Kimi K3, China's largest AI model at 2.8 trillion parameters. The company claims the model consistently outperforms OpenAI's GPT 5.5 and Anthropic's Claude Opus 4.8 on benchmarks including coding and general agents, though it still trails the leading-edge GPT 5.6 Sol and Claude Fable 5 in overall performance.
Cline v4.0.9 Adds GPT-5.6 ChatGPT Models, Fixes Token Count Over-Reporting
Cline, the AI coding assistant VS Code extension, released v4.0.9 on July 16, 2024, adding support for GPT-5.6 ChatGPT subscription models. The update fixes a bug where token counts were over-reported from OpenAI-compatible providers due to improper handling of cumulative usage snapshots.
Comments
Loading...