NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #1941 most downloaded on PyPI
cuda.core: pythonic CUDA module
Last release 11 days ago
23 Sep 2026
Ships fairly regularly
a new release about every 4 weeks
Nearly every release is documented
notes for 17 of 18 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
18 releases · first in 2024
One column per month.
https://nvidia.github.io/cuda-python/cuda-core/1.2.1/release/1.2.1-notes.html
https://nvidia.github.io/cuda-python/cuda-core/1.2.0/release/1.2.0-notes.html
Nothing published for this version
https://nvidia.github.io/cuda-python/cuda-core/1.1.0/release/1.1.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/1.0.1/release/1.0.1-notes.html
https://nvidia.github.io/cuda-python/cuda-core/1.0.0/release/1.0.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/0.7.0/release/0.7.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.6.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.5.1-notes.html
https://nvidia.github.io/cuda-python/cuda-core/0.5.0/release/0.5.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/0.4.2/release/0.4.2-notes.html
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.4.1-notes.html
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.4.0-notes.html
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.3.2-notes.html
cuda.core v0.3.2 release announcementAll functionalities are currently hosted under the cuda.core.experimental namespace. Once the features become stable they will be moved out of experimental.
cuda.core by @leofang in https://github.com/NVIDIA/cuda-python/pull/743SPDX-FileCopyrightText cleanup by @rwgk in https://github.com/NVIDIA/cuda-python/pull/753cuda.bindings to 13.0.0 by @leofang in https://github.com/NVIDIA/cuda-python/pull/792Device.set_current() faster by @leofang in https://github.com/NVIDIA/cuda-python/pull/781Full Changelog: https://github.com/NVIDIA/cuda-python/compare/cuda-core-v0.3.1...cuda-core-v0.3.2
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.3.1-notes.html
cuda.core v0.3.1 release announcementAll functionalities are currently hosted under the cuda.core.experimental namespace. Once the features become stable they will be moved out of experimental.
LegacyPinnedMemoryResource.allocate as kernel parameters by @shwina in https://github.com/NVIDIA/cuda-python/pull/717Full Changelog: https://github.com/NVIDIA/cuda-python/compare/cuda-core-v0.3.0...cuda-core-v0.3.1
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.3.0-notes.html
cuda.core v0.3.0 release announcementAll functionalities are currently hosted under the cuda.core.experimental namespace. Once the features become stable they will be moved out of experimental.
CUResult, cudaError explanations by @rwgk in https://github.com/NVIDIA/cuda-python/pull/503cuda.core license to Apache-2.0 & make contributing guides clear by @leofang in https://github.com/NVIDIA/cuda-python/pull/583ObjectCode instances by @brandon-b-miller in https://github.com/NVIDIA/cuda-python/pull/660ObjectCode constructors by @brandon-b-miller in https://github.com/NVIDIA/cuda-python/pull/652launch() + Fix fp16 bug by @leofang in https://github.com/NVIDIA/cuda-python/pull/669windows-2022 for Windows CI builds by @cryos in https://github.com/NVIDIA/cuda-python/pull/672ObjectCode to have a name by @leofang in https://github.com/NVIDIA/cuda-python/pull/682Device constructor by @leofang in https://github.com/NVIDIA/cuda-python/pull/460cuda.core to v0.3.0 by @leofang in https://github.com/NVIDIA/cuda-python/pull/703Full Changelog: https://github.com/NVIDIA/cuda-python/compare/cuda-core-v0.2.0...cuda-core-v0.3.0
Ensure deprecation warnings from cuda.bindings are swallowed by @ksimpson-work in https://github.com/NVIDIA/cuda-python/pull/404
cuda.core v0.2.0 release announcementAll functionalities are currently hosted under the cuda.core.experimental namespace. Once the features become stable they will be moved out of experimental.
ProgramOptions to facilitate the passing of runtime compile options to Program.Device and Kernel attributes.For full details please refer to the release note above.
stream attribute is removed from LaunchConfig. Instead, the Stream object should now be directly passed to launch as an argument.launch is changed by swapping positional arguments, the new signature is now (stream, config, kernel, *kernel_args)__cuda_stream__ from attribute to method.Program.compile method no longer accepts the options argument. Instead, you can optionally pass an instance of ProgramOptions to the constructor of Program.Device.properties now provides attribute getters instead of a dictionary interface..handle attribute of various cuda.core objects now returns the underlying Python object instead of a (type-erased) Python integer.jit_lto_fractal.py — Demonstrates just-in-time link-time optimization for fractal generation. (Device, LaunchConfig, Linker, LinkerOptions, Program, ProgramOptions) (#475)simple_multi_gpu_example.py — Example of using multiple GPUs. (Device, Program, LaunchConfig) (#304)show_device_properties.py — Displays detailed device properties. (Device) (#474)__cuda_stream__ from attribute to method by @NaderAlAwar in https://github.com/NVIDIA/cuda-python/pull/389cuda.bindings by @vzhurba01 in https://github.com/NVIDIA/cuda-python/pull/424ObjectCode as public API + prune unnecessary input arguments by @ksimpson-work in https://github.com/NVIDIA/cuda-python/pull/435dev.compute_capability by @leofang in https://github.com/NVIDIA/cuda-python/pull/459ObjectCode ptx constructor by @brandon-b-miller in https://github.com/NVIDIA/cuda-python/pull/470Linker example by @vzhurba01 in https://github.com/NVIDIA/cuda-python/pull/475__new__ approach to disabling __init__ by @rwgk in https://github.com/NVIDIA/cuda-python/pull/484cuda-core ver to v0.2.0 by @leofang in https://github.com/NVIDIA/cuda-python/pull/494test_timing() to avoid flaky tests. by @rwgk in https://github.com/NVIDIA/cuda-python/pull/498test_timing flakiness under Windows by @rwgk in https://github.com/NVIDIA/cuda-python/pull/508.decode() calls to .decode("utf-8", errors="backslashreplace") by @rwgk in https://github.com/NVIDIA/cuda-python/pull/510@functools.lru_cache decorator for get_binding_version() by @rwgk in https://github.com/NVIDIA/cuda-python/pull/512Full Changelog: https://github.com/NVIDIA/cuda-python/compare/cuda-core-v0.1.1...cuda-core-v0.2.0
https://nvidia.github.io/cuda-python/cuda-core/latest/release/0.1.1-notes.html
cuda.core v0.1.1 release announcementAll functionalities are currently hosted under the cuda.core.experimental namespace. Once the features become stable they will be moved out of experimental.
Linker for runtime linking (using nvJitLink or driver APIs)StridedMemoryView and @args_viewable_as_strided_memory to arbitrary Python objects that support either DLPack or CUDA Array InterfaceFor full details please refer to the release note above.
_util.device_from_ctx by @ksimpson-work in https://github.com/NVIDIA/cuda-python/pull/203__del__ with weakref.finalize() by @rwgk in https://github.com/NVIDIA/cuda-python/pull/246StridedMemoryView/@args_viewable_as_strided_memory by @leofang in https://github.com/NVIDIA/cuda-python/pull/247cuda.core version to v0.1.1 by @leofang in https://github.com/NVIDIA/cuda-python/pull/290cluster to LaunchConfig to support thread block clusters on Hopper by @leofang in https://github.com/NVIDIA/cuda-python/pull/261StridedMemoryView by deferring the check for whether a capsule is versioned by @leofang in https://github.com/NVIDIA/cuda-python/pull/292cuda.core v0.1.1 final doc touch by @leofang in https://github.com/NVIDIA/cuda-python/pull/301Full Changelog: https://github.com/NVIDIA/cuda-python/compare/cuda-core-v0.1.0...cuda-core-v0.1.1
Your coding agent can read these notes before it upgrades. Set up the MCP server →