NVIDIA Ising vs OpenMythos

NVIDIA Ising is the world's first family of open-source quantum AI models, launched April 14, 2026 on World Quantum Day. It targets two of the most expensive bottlenecks in making quantum processors useful: calibration (tuning the QPU to operate correctly) and error correction (detecting and fixing quantum errors in real-time). Both are currently handled by hand or with classical algorithms that don't scale. Ising Calibration is a 35-billion-parameter vision-language model fine-tuned to read experimental measurements from a quantum processing unit and infer the precise adjustments needed to tune it, reducing calibration time from days to hours when wrapped in an agentic loop. Ising Decoding ships two 3D convolutional neural network variants (0.9M and 1.8M parameters) for surface-code quantum error correction — up to 2.5× faster and 3× more accurate than pyMatching, the current open-source standard decoder. All models are available on GitHub, Hugging Face, and build.nvidia.com, alongside training data, workflows, and NVIDIA NIM microservices for fine-tuning on custom QPU hardware. Early adopters include Fermi National Accelerator Laboratory, Harvard, Lawrence Berkeley National Lab, IQM Quantum Computers, and the UK National Physical Laboratory. For quantum startups working to make NISQ devices practically useful, Ising dramatically reduces the engineering burden that today consumes much of their engineering bandwidth.

OpenMythos is a PyTorch reconstruction of the suspected architecture underlying Anthropic's Claude Mythos model, built entirely from published research. Creator Kye Gomez hypothesizes that Mythos uses a Recurrent-Depth Transformer (RDT) — where a subset of transformer layers loops multiple times per forward pass with shared weights rather than stacking unique layers. This allows the model to simulate "thinking" by iterating over the same compute graph, giving it emergent chain-of-thought behavior without explicit CoT prompting. At 770M parameters, the OpenMythos implementation reportedly matches the downstream quality of a 1.3B standard transformer on benchmarks. The architecture combines Multi-Latent Attention for memory compression, LTI (Linear Time-Invariant) stability constraints to prevent training instability during recurrence, Mixture of Experts routing for specialization, and Adaptive Computation Time (ACT) halting to decide when to stop looping per token. The project exploded on GitHub within days — 6.2k stars, 1.2k forks — and Kye's X announcement drove massive engagement (4.1k likes, 4.5k reposts). Community reaction is genuinely divided: AI researchers calling it "the most sophisticated reverse-engineering of an LLM architecture I've seen" while Anthropic has not confirmed or denied any of the architectural claims. This is an educated speculation backed by real engineering, not a marketing exercise.