Celeris-1 Launches Near-GPT-5 Intelligence at 15x Faster Speed
Celeris AI Labs, founded by ex-Marqo entrepreneurs, releases general-purpose language model promising frontier intelligence with reduced latency.
Celeris AI Labs announced Celeris-1, a general-purpose language model positioned as near-GPT-5 level intelligence with 15x faster responses than comparable frontier models. The startup was founded by Tom Hamer and Jn2clark, who previously launched Marqo, with research experience from Stanford, UCL, and Cambridge and engineering roles at Amazon Robotics. The launch was highlighted in retweets by AI figures including Tim Shi, an early OpenAI researcher and Cresta co-founder, and tech commentator Robert Scoble. No independent benchmarks or deployment details have been confirmed beyond the announcement.
Combined views
575
2 posts, first seen 4h ago
Reactions from ranked influencers
2 posts
Excited to launch Celeris-1 - a general purpose language model which delivers near-GPT-5 level intelligence with 15x faster response times. Celeris-1 delivers a p50 response latency of 157ms - around 15× faster than GPT-5-mini and 17× faster than GPT-5 - while scoring comparably 76% on MMLU-Pro, compared with 78% and 81%, respectively. The model is available starting today. Sign up at http://celeris.ai
Introducing celeris-1. A general purpose language model which delivers near-GPT-5 level intelligence with 15x faster response times. We’ve developed a new inference architecture that uses diffusion techniques instead of conventional autoregressive generation - unlocking dramatically better speed while maintaining frontier-level intelligence. Celeris-1 delivers a p50 response latency of 157ms - around 15× faster than GPT-5-mini and 17× faster than GPT-5 - while scoring comparably 76% on MMLU-Pro, compared with 78% and 81%, respectively. On tokens per second, Celeris-1 achieves a throughput of 1,280 tokens per second vs 144 for gemini-3.5 flash-light using a reconstructed version of @ArtificialAnlys's tokens per second benchmark dataset. The model is available starting today. Sign up at http://celeris.ai
Our portfolio founders @tom_w_hamer and @jn2clark, who previously launched Marqo, have announced @Celeris_ai Labs, an AI research lab with a mission to close the gap between frontier intelligence and real-time speed. Today they released their first model, Celeris-1. We backed Tom and Jesse because of our conviction in their rare combination of technical depth and grit. Between them, they bring deep research pedigree spanning Stanford, UCL and Cambridge, and hands-on experience building ML and core infrastructure at Amazon Robotics, AWS and Stitch Fix. We believed this was a team with the expertise and determination to tackle the largest opportunities in AI, and Celeris Labs is exactly that: a hard, foundational infrastructure problem of the kind that led us to them in the first place. Congratulations to Tom, Jesse and the research team for the long hard cycles that have gone into this! In AI, speed is everything and the benchmarks from Celeris Labs are exceptional. Check them out: http://celeris.ai
Excited to launch Celeris-1 - a general purpose language model which delivers near-GPT-5 level intelligence with 15x faster response times. Celeris-1 delivers a p50 response latency of 157ms - around 15× faster than GPT-5-mini and 17× faster than GPT-5 - while scoring comparably 76% on MMLU-Pro, compared with 78% and 81%, respectively. The model is available starting today. Sign up at http://celeris.ai