AI News (2026/9/10): ChatGPT Voice Mode Supports GPT-5.6 Sol and GPT-6 Astra Invocation
Executive Summary:
OpenAI launched an upgraded voice mode solution for ChatGPT on September 10. This version allows users to actively select GPT-5.6 Sol or GPT-6 Astra models for complex reasoning tasks during conversations, introducing dynamic model switching mechanisms and a unified billing system.
[ChatGPT Voice Mode Adds Model Selection to Transform Interaction Efficiency]
OpenAI has released an upgraded voice mode solution for ChatGPT. This version allows users to actively select GPT-5.6 Sol or GPT-6 Astra models for complex reasoning tasks during conversations, introducing dynamic model switching mechanisms and a unified billing system.
News Details
This update optimizes model scheduling logic specifically for voice interaction scenarios. When users pose questions involving search requirements or deep analysis needs, the system will automatically activate designated underlying text models according to predefined rules. All cross-modal handover operations are now included in the message quota calculation system for Pro user subscriptions.
Core Highlights
[Dynamic Model Selection]: Pro users can switch between GPT-5.6 Sol (128k context window) or GPT-6 Astra (256k context window) models in voice conversation interfaces using the
/modelcommand[Intelligent Task Routing]: The system automatically identifies task types requiring underlying text model invocation (e.g., multi-document retrieval, mathematical calculations) through semantic parsing, reducing average response latency to 1.2 seconds
[Unified Billing System]: Each cross-modal handover is billed as a complete message unit (including speech-to-text, model reasoning, and text-to-speech phases), consuming 3.8 times the message quota of standard voice interactions
AI-ALL In-depth Analysis
This update marks the entry of multimodal interaction into a precision scheduling era. By exposing model selection interfaces at the infrastructure level, OpenAI has deeply integrated reasoning capabilities in voice scenarios with its text ecosystem, establishing a new value assessment framework while enhancing interaction efficiency.
Developers should focus on two critical changes:
- Increased Context Management Complexity: The 256k window of GPT-6 Astra necessitates architectural redesign of conversation history storage strategies at the application layer
- Heightened Cost Control Challenges: The cross-modal task routing mechanism causes API invocation costs to exhibit nonlinear growth characteristics
The update will create three major industry impacts:
- Accelerate RAG technology deployment in voice scenarios
- Speed up Agentic Workflow adoption in consumer-grade products
- Establish new pricing benchmarks for multimodal interaction services
The current implementation still has optimization potential for context fragmentation issues in continuous dialogue scenarios and long-term task processing.
