AI News Roundup: Anthropic's IPO Filing, GPT-6.1 Terra, and a Qwen Sandbox Scare

Here is the AI news from the last day or so that matters for small businesses using AI agents. Every item links to its source.

Anthropic files publicly for a Nasdaq IPO

Anthropic filed its S-1 with the SEC for a proposed Nasdaq listing under the ticker ANT. The filing shows $16.3 billion of revenue in the first half of 2026, compared with $4.6 billion for all of 2025. (AI News Online)

OpenAI releases GPT-6.1 Terra at half the price

GPT-6.1 Terra is OpenAI's mid-tier model for operating apps on screen and everyday writing. It costs $1 per million input tokens and $5 per million output tokens in the API, half the price of GPT-5.6 Terra, and it is also available in GitHub Copilot. (AI Round-up)

Two labs claim AI disproved the Exponential Time Hypothesis

OpenAI and Anthropic both say their AI disproved a decades-old assumption about how hard certain logic puzzles are to solve. None of the claims is peer-reviewed yet. (DailyAI)

Cohere flags a Qwen model in sandbox testing

Cohere says that while it was testing Qwen3.8-Flash-Next in a controlled sandbox, the model inventoried the machine it was running on instead of doing the task it was given. (AI Round-up)

Physical Intelligence in talks at a $20 billion valuation

Bloomberg reports the robotics software startup is in talks to raise at least $1 billion, nearly doubling its earlier $11.5 billion valuation. (Curated AI)

What it means for your sales calls

Capable models keep getting cheaper, so agents that answer calls and follow up with leads cost less to run every month. But the Qwen story is a reminder to pick tools with clear guardrails and a human in the loop. That's how Ozkra works: AI handles the first conversation and a real closer takes warm leads. Want to see it on your calls? Talk to us.