NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI
PyPI, #3598 most downloaded on PyPI
A framework for machine learning on Apple silicon.
Latest stable
0.32.4
Last release 1 months ago
14 Oct 2026
Ships fairly regularly
a new release about every 4 weeks
Some releases are documented
notes for 32 of 60 stable releases
Nothing withdrawn
no release was ever pulled
3 years old
60 releases · first in 2023
Drag across the ruler to compare, or tap a release to open it.
Adding VJP for the scaled dot product attention. by @tpegolotti in https://github.com/ml-explore/mlx/pull/4563
unique operation by @louen in https://github.com/ml-explore/mlx/pull/4501col_reduce_longcolumn for negative-stride views by @mateuuszzzzz in https://github.com/ml-explore/mlx/pull/4529mx.block_masked_mm by @michalk8 in https://github.com/ml-explore/mlx/pull/4652Full Changelog: https://github.com/ml-explore/mlx/compare/v0.32.3...v0.32.4
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Fix scan and sort for a zero-size axis by @eyupcanakman in https://github.com/ml-explore/mlx/pull/4340
linear mode align_corners=False by @sp4s-s in https://github.com/ml-explore/mlx/pull/4500Full Changelog: https://github.com/ml-explore/mlx/compare/v0.32.2...v0.32.3
Preserve subnormal float values when casting to bool by @reckylurker in https://github.com/ml-explore/mlx/pull/4224
RMSNormVJP backward writes a full {n_rows, D} gw_temp intermediate by @JasonHonKL in https://github.com/ml-explore/mlx/pull/4293Full Changelog: https://github.com/ml-explore/mlx/compare/v0.32.1...v0.32.2
Fix int64 type cast error when loading GGUF metadata arrays by @danlee2002 in https://github.com/ml-explore/mlx/pull/3823
thread address space qualifier becoming explicit in metal 4.1 by @louen in https://github.com/ml-explore/mlx/pull/3963precise argument by @apocryphx in https://github.com/ml-explore/mlx/pull/4178Full Changelog: https://github.com/ml-explore/mlx/compare/v0.32.0...v0.32.1
Generate qmm implementaions with cmake by @zcbenz in https://github.com/ml-explore/mlx/pull/3424
copy keyword to mx.asarray by @eyupcanakman in https://github.com/ml-explore/mlx/pull/3510Full Changelog: https://github.com/ml-explore/mlx/compare/v0.31.2...v0.32.0
Wider support for cuda quantized matmuls (#3352, #3268, #3321, #3417, #3255)
__index__ by @aisk in https://github.com/ml-explore/mlx/pull/3210metal as the front-end for the metal linker by @louen in https://github.com/ml-explore/mlx/pull/3354Full Changelog: https://github.com/ml-explore/mlx/compare/v0.31.1...v0.31.2
Bump the patch version by @angeloskath in https://github.com/ml-explore/mlx/pull/3185
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.31.0...v0.31.1
Initial version of QMMs for CUDA
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.6...v0.31.0
Much faster bandwidth with JACCL on macOS >= 26.3 (some numbers)
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.5...v0.30.6
patch by @awni in https://github.com/ml-explore/mlx/pull/3074
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.4...v0.30.5
Metal: Much faster vector fused grouped-query attention for long context
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.3...v0.30.4
Support nvfp4 and mxfp8 quantized ops on Metal
mlx.nn.init.he_normal and mlx.nn.hard_tanh by @Redempt1onzzZZ in https://github.com/ml-explore/mlx/pull/2968grid_dim_x calculations by @CC-Yeh in https://github.com/ml-explore/mlx/pull/2980Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.1...v0.30.3
RDMA over thunderbolt with the JACCL backend (macOS >= 26.2) (some numbers)
mx.core.load type annotation by @CC-Yeh in https://github.com/ml-explore/mlx/pull/2819mx.core.linspace type annotation by @CC-Yeh in https://github.com/ml-explore/mlx/pull/2820Numpy interfaces for masked_scatter by @CC-Yeh in https://github.com/ml-explore/mlx/pull/2832Full Changelog: https://github.com/ml-explore/mlx/compare/v0.30.0...v0.30.1
Support for Neural Accelerators on M5 (macOS >= 26.2)
mx.depends to Python by @awni in https://github.com/ml-explore/mlx/pull/2606compile_commands.json by @andportnoy in https://github.com/ml-explore/mlx/pull/2645mx.median op by @awni in https://github.com/ml-explore/mlx/pull/2705zeros/ones_like by @CC-Yeh in https://github.com/ml-explore/mlx/pull/2726Full Changelog: https://github.com/ml-explore/mlx/compare/v0.29.0...v0.30.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Support for mxfp4 quantization (Metal, CPU)
mxfp4 quantization (Metal, CPU)mx.distributed supports NCCL back-end for CUDAcudaMemAdvise and cudaGraphAddDependencies for CUDA 13 by @andportnoy in https://github.com/ml-explore/mlx/pull/2525Path by @awni in https://github.com/ml-explore/mlx/pull/2543Full Changelog: https://github.com/ml-explore/mlx/compare/v0.28.0...v0.29.0
First version of fused sdpa vector for CUDA
Full Changelog: https://github.com/ml-explore/mlx/compare/v0.27.1...v0.28.0
Initial PyPi release of the CUDA back-end.
update and update_modules by @awni in https://github.com/ml-explore/mlx/pull/2239update_modules() when providing a subset by @angeloskath in https://github.com/ml-explore/mlx/pull/2308Full Changelog: https://github.com/ml-explore/mlx/compare/v0.26.0...v0.27.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Peaking at 23.89 TFlops on M2 Ultra benchmarks
mx.linalg.eigh and mx.linalg.eigvalshmx.nn.init.sparsemx.cumprod, mx.cumsummx.random.uniform and mx.random.bernoullimx.vmap with gather and constant outputsmx.array.__format__ with specNothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Block sparse matrix multiply speeds up MoEs by >2x
mx.bitwise_[or|and|xor], mx.[left|right]_shift, operator overloadsmx.metal.device_info to get better informed memory limitsmlx.optimizers.clip_grad_norm and mlx.utils.tree_reduce addedmx.arctan2mx.sqrt(2)Nothing published for this version
Nothing published for this version
Improvements for LLM generation
mx.async_evalmx.metal.start_capture and mx.metal.stop_capture for GPU debug/profilemx.expm1mx.stdmx.meshgridmx.random.multivariate_normalmx.cumsum (and other scans) for bfloatnn.upsample support bicubic interpolationNothing published for this version
Nothing published for this version
Perf improvements for attention ops:
mx.linalg.svd (CPU only)nn.RNN, nn.LSTM, nn.GRUFaster quantized matrix-vector multiplies
mx.fast.scaled_dot_product_attention fused opmx.fast.scaled_dot_product_attention fused opmx.arraymx.topkNothing published for this version
Default shapeless compilation for all activations
mx.compile(function, shapeless=True)mx.atleast_1d, mx.atleast_2d, mx.atleast_3dtolist with bfloat16 and float16argmax on M3Custom mx.fast.rope up to 20x faster
mx.fast subpackagemx.fast.rope up to 20x fastersafetensorsbfloat16 quantizated matrix-vector multipliesmx.fast subpackage with a fast RoPEmx.stream to set the default deviceoptimizers.step_decayoptimizers.cosine_decayopimtizers.exponential_decaySome functions are up to 10x faster (benchmarks)
mx.compile makes stuff go fast
mx.compile function transformation__abs__ overload for abs on arraysloc and scale in parameter for mx.random.normalmx.var to give inf with doff >= nelemnn.SequentialGradient checkpointing for training with mx.checkpoint
mx.checkpointmx.checkpointmx.linalg.qrmx.evalmx.diag, mx.diagonalarray.shape is a Python tupleint64 and uint64 reductionssum, prod, max, min, all, anyargmax, argmininf work, and fix mx.isinfmx.fullNaN in some binary ops
mx.logaddexp, mx.maximum, mx.minimummx.log1p with inf inputNative quantizations Q4_0, Q4_1, and Q8_0
Q4_0, Q4_1, and Q8_0Q4_0, Q4_1, and Q8_0)Module.save_weights supports safetensorsnn.init package with several commonly used neural network initializersAdafactor in nn.optimizersisinf and friends for integer typesint64, uint, and float320 inputsinf reads in gemvmx.arange crashes on NaN inputsFaster matmul: up to 2.5x faster for certain sizes, benchmarks
mx.isnan, mx.isinf, isposinf, isneginfmx.tilescatter_min and scatter_maxmx.eyemx.round to follow NumPy which rounds to evenInitial (and experimental) GGUF support
at[] syntax for scatter style operations: x.at[idx].add(y), (min, max, prod, etc)mx.array([x, y]))mx.inner, mx.outer+=, *=, -=, ...)mx.pi, mx.inf, mx.newaxis, …)cosine_similarity lossRoPE and ALiBitriretain_graphSupport for loading and saving HuggingFace's safetensor format
mlx.core.linalg sub-package with mx.linalg.norm (Frobenius, infininty, p-norms)tensordot and repeatBilinear,Identity, InstanceNormDropout2D, Dropout3DTransformer (pre/post norm, dropout)SoftSign, Softmax, HardSwish, LogSoftmaxRoPE positional encodingshinge, huber, log_coshquantize, dequantize, quantized_matmul
QuantizedLinear, ALiBi positional encodingsCore ops remainder, eye, identity
remainder, eye, identitymlx.nn
Nothing published for this version
Nothing published for this version
Nothing published for this version
One column per month.