The Arena

GPT-5.6 Sol powers OpenAI's revenue rebound over Anthropic

The pendulum in the OpenAI–Anthropic revenue race just swung back. Three weeks after Anthropic briefly passed OpenAI on quarterly revenue for the first time, OpenAI says GPT-5.6 Sol has it growing again — and the spending data behind the scenes backs that up.

OpenAI says GPT-5.6 Sol, launched July 9, has driven a sharp revenue turnaround — up 35 percent this quarter, with enterprise revenue climbing more than 50 percent. The numbers come from OpenAI's own commentary to CNBC, but the more interesting signal is independent: Ramp's business-API spending index shows OpenAI outgrowing Anthropic again in Q3, 82 percent quarter-over-quarter versus 76 percent. That reverses the trend we covered in mid-August, when Anthropic's $65 billion annualized run rate and $11.6 billion quarterly revenue put it ahead of OpenAI's $6.7 billion for the first time.

The reversal is worth taking seriously because it's not just OpenAI's word. Ramp's card-data index is one of the few real-time, third-party reads on which lab enterprises are actually paying, and it had shown Anthropic pulling ahead on momentum. A 6-point gap the other way in a single quarter is the kind of swing that suggests GPT-5.6 Sol landed with paying customers in a way its predecessors did not. Enterprise growth north of 50 percent is the figure to watch — that is the segment both labs are fighting hardest to win, and it is where margins live.

What happens next is the open question. The Decoder reports OpenAI's next model, Astra, is set to launch within weeks, and rumors point to Anthropic countering with an improved Fable 5.1. Both labs have also seen growth slow as open-weight models eat into the paid frontier tier, so a single quarter's swing is not a coronation. But the narrative that Anthropic had definitively taken the lead now has a competitor's comeback story attached to it — and the spending data says the comeback is real, at least for now.

What to watch: whether Astra's launch extends OpenAI's momentum or just borrows against it, and whether Anthropic's Fable 5.1 refresh closes the Q3 gap back.

Is this OpenAI's durable lead or a one-quarter bounce before the next model drop — tell us in the comments.

Sources: The Decoder · Ramp AI Index · CNBC · Bloomberg

Unitree's founder gives robots a two- to ten-year path to a 'ChatGPT moment'

China's robot sector spent this week at the World Robot Conference in Beijing celebrating its hottest-ever stretch — Unitree's Shanghai debut soared nearly sixfold before cooling — yet the industry's own leader is keeping expectations firmly in check. Unitree founder and CEO Wang Xingxing told

Galaxea preps G0.5 MAX and a new robot as it chases 10,000-unit deliveries

At the World Robot Conference in Beijing, Galaxea AI (星海图) laid out the most concrete plan yet for turning embodied-AI models into an actual production system — and put numbers on it. Chief scientist Hang Zhao said the company's G0.5 foundation model beat industry SOTA across eight

Open Source Radar — August 22: swarms, workspaces & observability

Today's board is the agent operating layer: a meta-harness for orchestrating teams of agents, Apache's new local-first agent workspace, and an analytics platform that wants to be every AI team's eyes on what agents are doing. Three projects, all fresh on daily

World models that ignore what people believe predict the wrong actions

A new "Mental World Modeling" framework adds beliefs, intentions, and social norms to world models — and even weak language models running it beat frontier giants answering directly. Current world models simulate physics but skip the people in the scene — and a new paper argues that omission makes them

Read more

Unitree's founder gives robots a two- to ten-year path to a 'ChatGPT moment'

Galaxea preps G0.5 MAX and a new robot as it chases 10,000-unit deliveries

Open Source Radar — August 22: swarms, workspaces & observability

World models that ignore what people believe predict the wrong actions