Urgent.News

What's breaking now, across thousands of outlets.

AI

OpenAI launches cheaper AI model, after scrapping one deemed unsafe

SAN FRANCISCO — OpenAI launched a new mid-range artificial intelligence model Tuesday that costs five times less than its flagship offering, one day after scrapping an upgraded version deemed unreliable. GPT-6.1 Sol ranks below OpenAI's most powerful model, GPT-6 Astra, released in early September. The company on Monday said it abandoned plans for GPT-6.1 Astra because the model frequently…

OpenAI launches cheaper AI model, after scrapping one deemed unsafe

OpenAI has unveiled GPT-6.1 Sol, a new model that boasts Astra-like performance at a fraction of the cost. According to the company, GPT-6.1 Sol performs nearly as well as Astra across numerous benchmarks, closely matching six instances of Astra’s capabilities. Moreover, GPT-6.1 Sol outperforms GPT-6 Sol in terms of factual accuracy.

The announcement comes during OpenAI's DevDay event, where the company is presenting updates on its latest model and service developments. However, GPT-6.1 Sol is not without its issues, as Astra, a rival model, is facing setbacks due to misbehavior.

Despite the challenges, OpenAI remains focused on advancing its AI capabilities. Today marks the debut of GPT-6.1 Sol, the first model in the 6.1-series to make a public appearance. While Astra continues to lead in terms of raw performance, OpenAI claims that GPT-6.1 Sol is rapidly closing the gap in terms of "intelligence," achieving near parity across multiple metrics with the flagship model.

Written by urgent.news from Android Authority's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at koreatimes.co.kr →

More in AI

Built a customer-support AI agent that actually remembers customers

Built a customer-support AI agent that actually remembers customers One problem I noticed with AI support agents is that they often make customers repeat the same information in every new…

  • SupportMemory AI agent retains customer information
  • Utilizes Hindsight persistent memory for recall and retention
  • Demonstrates improved customer support through context awareness

Your AI coding agent hit a limit? Keep working with MateMCP

A frustrating thing kept happening during real coding work: I would be deep into a task, the built-in coding/agent mode would hit its usage limit, and the agent would stop — while the conversation…

  • MateMCP offers alternative execution path for AI conversations.
  • Provides shell commands, file access, and project workspaces.
  • Maintains conversation usefulness despite AI provider's usage limits.

Building an AI-Powered Incident Response Agent with Hindsight

Hack With Hyderabad 3.0 | AI Agents | DevOps | Hindsight Memory 🚨 When Production Breaks, What If Your AI Agent Could Remember? Imagine this. It's 2 AM.

  • AI-powered Incident Response Agent with memory called Hindsight
  • Agent recalls past incidents to provide context during investigations
  • Agent recommends actions and learns from incident experiences

More from Tuesday 29 September →