doubleAI unveils Maxwell, a runtime model that rewrites running binaries
Original title: Maxwell: A new kind of model that sits between code and hardware
doubleAI says Maxwell proves each rewrite equivalent to the original and claims large speedups on GPUs and CPUs.
doubleAI introduced Maxwell, described as a code execution model that watches a program while it runs, rewrites the binary for that machine and data, and formally proves each change equivalent to the original. The company says it runs on CPUs and GPUs and speeds data pipelines and AI inference kernels. Published demos compare Whisper-small and NetworkX runs with and without Maxwell, and doubleAI cites customer and benchmark results that it has not independently confirmed here.
Unconfirmed: Speedups and benchmark wins are doubleAI’s own claims, not independently verified in this item.
- Vendor claims 38× faster data pipelines and 4.25× faster inference kernels.
- LTX-2.5 4K generation is claimed at 134 seconds versus 250 seconds on two B200s.
- Whisper-small demo: 106.23 seconds without Maxwell, 49.82 seconds with it.
- NetworkX betweenness demo: 61.44 seconds versus 17.50 seconds, identical results.
- doubleAI says Maxwell beat PyTorch baselines on 90% of SOL-ExecBench Blackwell kernels.
- Organizations
- LTX
- AI models
- MaxwellLTX-2.5Whisper-small
From the original article
2 min read at heymaxwell.ai
Maxwell understands and improves code while it runs, reasoning about hardware and data, not just code. Every improvement is formally proven equivalent to the original code. Read the manifesto A short animated film, with narration and captions, introducing Maxwell: most code never sees the hardware it runs on or the data it processes.
By laginimaineb · heymaxwell.ai
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.