Celeris Releases Celeris-1 Diffusion Language Model
Celeris has released Celeris-1, a diffusion language model available through an API compatible with OpenAI, the company announced in a press release. The company says Celeris-1 delivers near GPT-5 level intelligence with response times up to 15 times faster than standard autoregressive models.
Celeris-1 uses diffusion for language generation. Instead of producing one token at a time, the model starts with a rough version of a full response and refines it over several passes.
The company says Celeris-1 generates 1,664 tokens per second, more than 5 times the output speed of Inception's Mercury 2, and scored 75.9 percent on MMLU-Pro. It is described as the second commercially available diffusion language model API after Mercury.
Celeris says the model is built for structured workloads that need fast responses, including agent tool orchestration, classification, data extraction, voice pipelines, live translation, and session assistants. Developers can access Celeris-1 by updating their API base URL and key through celeris.ai.
We hope you enjoyed this article.
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Subscribe to Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates
Market report
.png)
2025 Generative AI in Professional Services Report
This report by Thomson Reuters explores the integration and impact of generative AI technologies, such as ChatGPT and Microsoft Copilot, within the professional services sector. It highlights the growing adoption of GenAI tools across industries like legal, tax, accounting, and government, and discusses the challenges and opportunities these technologies present. The report also examines professionals' perceptions of GenAI and the need for strategic integration to maximize its value.
Read more