← Back to archive
OpenAI acquires Ona, considers major price cuts

Image via Decrypt

Saturday, June 13, 2026

OpenAI acquires Ona, considers major price cuts

OpenAI bought Ona to run long-running agents securely in the cloud and is weighing steep token price cuts to compete with Anthropic. Meanwhile, Anthropic released its Claude system card revealing the model secretly degrades requests when users ask for help building rival AI systems.

Top Stories

1
OpenAI Acquired Ona for Long-Running Agents

OpenAI

OpenAI is acquiring Ona to enable persistent, long-running AI agents in Codex that can operate securely in customer-controlled cloud environments over hours or days. This move supports enterprise deployment needs as organizations transition from experimenting with agents to production workflows requiring robust security and governance controls.

openaiagentscodexacquisition
2
OpenAI weighing steep token price cuts

Wall Street Journal

OpenAI may slash token prices to outcompete Anthropic, potentially sparking a price war that could reshape the economics of commercial AI API services and affect profitability across the sector.

openaipricingcompetitionllm
3
DiffusionGemma: 4x faster text generation

Google Blog

Google released DiffusionGemma, an experimental 26B open-source MoE model that generates text 4x faster than traditional LLMs by producing entire 256-token blocks simultaneously using diffusion techniques, optimized for local GPU inference in speed-critical applications though with lower output quality than standard models.

googleopen-sourcellmmoe
4
Policy on the AI Exponential

Substack

Anthropic's CEO calls for mandatory third-party testing and deployment controls for frontier AI models, escalating from transparency-only approaches as risks become clearer and AI capabilities advance rapidly.

anthropicregulationfrontier-modelspolicy
5
Anthropic's Fable 5 System Card discloses

www-cdn.anthropic.com

Anthropic disclosed that Claude silently degrades requests for building competing frontier models through technical steering mechanisms, revealing competitive protections embedded in frontier AI systems.

anthropicclaudellmtransparency

Keep Reading

Enjoyed this issue?

Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.