Celeris benchmarks celeris-1 on GSM8K, IFEval, and DROP
The model achieves 85.3% average accuracy with median latency of 99 ms, close to GPT-5 performance.
See the latest news and media coverage for Celeris. We track all announcements, press releases, and industry mentions in real time, all in one place.
AI research lab
celeris.aiLast updated
The model achieves 85.3% average accuracy with median latency of 99 ms, close to GPT-5 performance.
It uses diffusion inference for near-GPT-5 intelligence at 15x faster speeds
Celeris-1 delivers 1,596 tok/s, 8% faster than Cerebras and 3.5x faster than Groq in end-to-end benchmarks.
On a standardized workload, it generated 1,664 tokens per second, 5.1 times faster than the next fastest API.
Celeris-1 is a diffusion LLM that delivers near-GPT-5 intelligence with up to 15x faster response times. SAN FRANCISCO, July 27, 2026 /PRNewswire/ -- Celeris, an...
Track Celeris and your other target companies to get real-time alerts and weekly summaries delivered straight to your inbox.
Browse news for competitors to Celeris and other trending companies.