MAX Nightly 26.8.0.dev2026102914 (Mojo 1.3.0.dev2026102914) Released

[Modular]

A new nightly version has been released!

See the quickstart guide for installation instructions: Quickstart | MAX

MAX changelog updates:

• FLUX.2 models now generate more than one image per request, so
  num_images above 1 in max serve and num_images_per_prompt in
  generate_images() and AutoModel.generate() work for them. Such a
  request used to fail in the first denoising step.

• Serving metric names now use max.* instead of maxserve.*. Batch
  instruments use max.batch.*, such as max.batch.size. Prometheus
  exposes the corresponding max_* families. Update collector allowlists,
  dashboards, and alerts with the deployment; existing metric history stays
  under its previous names. Units, labels, histogram boundaries, and
  benchmark result fields are unchanged.

• max serve now rejects an image generation request that asks for more
  than 4 images with a 422 error naming num_images, the same response as
  other out-of-range image options. generate_images() and
  AutoModel.generate() raise a ValueError for num_images_per_prompt
  above 4 before any prompt runs. A request had no upper limit before.

• The eager Module framework moved from max.experimental.nn to max.nn,
  and the graph-based layers that max.nn held moved to max.nn.legacy.
  max.nn.legacy is deprecated and will be removed in MAX 27.1. Code that
  uses it gets a DeprecationWarning naming the line that used it. To keep
  using the graph-based layers until then, change from max.nn import ... to
  from max.nn.legacy import ....

• max.nn.legacy.Transformer, DistributedTransformer,
  logits_postprocess() and distributed_logits_postprocess() return a
  max.nn.legacy.graph_groups.DecoderOutputs of named groups instead of a
  tuple.
  max.functional’s matmul() require both inputs to share a dtype, which
  the result keeps, so max.nn.legacy.Linear raises when its input and weight
  dtypes differ. max.nn.Linear casts its input to the weight’s dtype, so a

• Added max.driver.Buffer.memset(value) to fill a contiguous buffer or
  slice in place with a byte value, without allocating a source buffer.

  ops.resize_linear, ops.resize_nearest, and their max.functional
  counterparts take them for coordinate_transform_mode and

• Removed max.nn.legacy.FusedMLP. Use max.nn.legacy.MLP, which takes the
  separate gate_proj and up_proj weights a checkpoint ships and fuses their
  matmuls in the graph.

• Removed the per_layer_buffers KV cache option, which gave each layer its
  own buffer to get past a device’s per-allocation size limit, along with the
  kv_blocks_per_layer and kv_scales_per_layer fields of
  KVCacheInputsPerDevice. A KV cache is again one buffer per device, capped
  at that limit.

• Enabled automatic fusion of eligible bfloat16 matrix multiplications and
  SwiGLU activations on NVIDIA B200 GPUs.

• Added DeviceContext.device(), which returns a Device view of the
  accelerator with an id field and get_name(), get_attribute(),
  is_host_unified(), get_arch() and can_access_peer() methods.
  DeviceContext.id(), name(), get_attribute(), is_host_unified() and
  can_access() are deprecated: write ctx.device().id,
  ctx.device().get_name(), ctx.device().can_access_peer(peer.device())
  and so on.

• The free memory that CPU and Apple GPU devices report now accounts for
  memory used by other processes. Before, MAX could start a model that didn’t
  fit in memory, which on macOS could exhaust swap and crash the machine.

Mojo changelog updates:

• [External] [Stdlib] Add os.path.normpath (#107338)
• [External] [mojo][stdlib] Re-enable ConvertibleToPython conformance for builtin types (#106687)
• [stdlib] Remove the Powable, CeilDivableRaising, and _Expable traits
• [stdlib] Make FileHandle own its descriptor through FileDescriptor

Raw MAX diff: https://github.com/modular/modular/compare/cb987e0e4f4df28d19f04dd586dc76713751cf65...8ed65fc6ac15950ce62da1f8e8d50c2c7c528065
Current Mojo changelog: https://github.com/modular/modular/blob/main/Mojo/docs/site/nightly-changelog.md