# OpenAI lost U.S. GPT trademark

*Science. 58% certain. Thursday, 1 October 2026*

> Speculative fiction assembled from prediction-market prices. This has not happened.

**OpenAI kept its Code Arena lead while falling short on a U.S. GPT trademark and a new Millennium Prize solution.**

OpenAI experienced a mixed set of outcomes as of early October 2026. The company retained the top spot in the Code Arena WebDev competition. It placed Second in the AI Agent Lab rankings released at the end of September. It also classified a model as Critical, lost a U.S. GPT trademark, and Did not announce a new Millennium Prize solution.

OpenAI's WebDev model emerged as the best performer in the Code Arena contest that concluded in September. The assessment compared models from Anthropic, Alibaba and SpaceXAI, with OpenAI maintaining the lead. The result reinforced the company's position in web development AI tools.

In the AI Agent Lab evaluation, OpenAI secured the Second-best ranking. The list, compiled at the end of September, placed another participant ahead of OpenAI. The standing highlighted strong performance but fell short of the top slot.

OpenAI designated one of its models as having Critical capability under its Preparedness Framework before 1 October. The classification signals that the model poses heightened safety considerations. The move aligns with the company's broader risk-management strategy.

The United States Patent and trademark Office denied OpenAI a trademark for the term "GPT". Moreover, OpenAI Did not announce any new solution to the Millennium Prize problems by the end of September. The organization also Did not solve the hodge?? Challenge, confirming the absence of that breakthrough.

## Analysis

The combination of technical successes and setbacks illustrates the multifaceted pressures facing AI developers. Holding the Code Arena lead and a high AI Agent ranking shows continued strength in product performance, while the loss of a trademark and the absence of a Prize-winning breakthrough expose vulnerabilities in legal and research domains.

Without exclusive rights, competitors might adopt similar naming, potentially diluting brand distinctiveness and complicating trademark enforcement strategies.

Classifying a model as Critical under the Preparedness Framework signals that OpenAI is actively managing higher-risk capabilities. This designation likely triggers enhanced oversight, testing, and possibly usage restrictions, reflecting a precautionary approach to emerging safety concerns within advanced AI systems.

The company must balance continued innovation with legal defenses and safety protocols, while the lack of a new Millennium Prize solution leaves open the question of whether further breakthroughs are forthcoming.

## What this was built from

- **Which company has the best Code Arena WebDev AI model end of September?**: OpenAI, 78%; settles Thursday, 1 October 2026 (polymarket). [Market](https://polymarket.com/event/which-company-has-the-best-code-arena-webdev-ai-model-end-of-september-20260717140116512)
- **Did OpenAI just solve hodge??**: No, 58%; settles Thursday, 1 October 2026 (manifold). [Market](https://manifold.markets/Incompleteusern/did-openai-just-solve-hodge)
- **Will OpenAI be granted a trademark on "GPT" (in the U.S.)?**: No, 89%; settles Thursday, 1 October 2026 (manifold). [Market](https://manifold.markets/ScottLawrence/will-openai-be-granted-a-trademark)
- **Second-best AI Agent Lab end of September?**: OpenAI, 88%; settles Thursday, 1 October 2026 (polymarket). [Market](https://polymarket.com/event/second-best-ai-agent-lab-end-of-september-20260812152609626)
- **Will OpenAI announce another Millennium Prize solution by September 30, 2026?**: No, 73%; settles Thursday, 1 October 2026 (polymarket). [Market](https://polymarket.com/event/openai-announces-another-millennium-prize-solution-by)
- **Will OpenAI classify any model as "Critical" capability under its Preparedness Framework before Oct 1, 2026?**: Yes, 96%; settles Thursday, 1 October 2026 (manifold). [Market](https://manifold.markets/Terminator2/will-openai-classify-any-model-as-c)