This change adds support for qwen3.5-next-moe models (qwen3-next/qwen3.5-next/qwen3-coder) to the MLX runner. It also:
* introduces recurrent cache support and related MLX ops
* updates pipeline/runner integration and adds tests
* properly quantizes stacked expert tensors
* a Gated Delta Metal kernel for fast SSM inference
* adds new MLX calls for Conv1d, DepthwideConv1d, Contiguous, Exp, Log, SoftmaxAxis
* don't require pulling stubs for cloud models
This is a first in a series of PRs that will better integrate Ollama's
cloud into the API and CLI. Previously we used to have a layer of
indirection where you'd first have to pull a "stub" model that contains
a reference to a cloud model. With this change, you don't have to pull
first, you can just use a cloud model in various routes like `/api/chat`
and `/api/show`. This change respects
<https://github.com/ollama/ollama/pull/14221>, so if cloud is disabled,
these models won't be accessible.
There's also a new, simpler pass-through proxy that doesn't convert the
requests ahead of hitting the cloud models, which they themselves
already support various formats (e.g., `v1/chat/completions` or Open
Responses, etc.). This will help prevent issues caused by double
converting (e.g., `v1/chat/completions` converted to `api/chat` on the
client, then calling cloud and converting back to a
`v1/chat/completions` response instead of the cloud model handling the
original `v1/chat/completions` request first).
There's now a notion of "source tags", which can be mixed with existing
tags. So instead of having different formats like`gpt-oss:20b-cloud` vs.
`kimi-k2.5:cloud` (`-cloud` suffix vs. `:cloud`), you can now specify
cloud by simply appending `:cloud`. This PR doesn't change model
resolution yet, but sets us up to allow for things like omitting the
non-source tag, which would make something like `ollama run
gpt-oss:cloud` work the same way that `ollama run gpt-oss` already works
today.
More detailed changes:
- Added a shared model selector parser in `types/modelselector`:
- supports `:cloud` and `:local`
- accepts source tags in any position
- supports legacy `:<tag>-cloud`
- rejects conflicting source tags
- Integrated selector handling across server inference/show routes:
- `GenerateHandler`, `ChatHandler`, `EmbedHandler`,
`EmbeddingsHandler`, `ShowHandler`
- Added explicit-cloud passthrough proxy for ollama.com:
- same-endpoint forwarding for `/api/*`, `/v1/*`, and `/v1/messages`
- normalizes `model` (and `name` for `/api/show`) before forwarding
- forwards request headers except hop-by-hop/proxy-managed headers
- uses bounded response-header timeout
- handles auth failures in a friendly way
- Preserved cloud-disable behavior (`OLLAMA_NO_CLOUD`)
- Updated create flow to support `FROM ...:cloud` model sources (though
this flow uses the legacy proxy still, supporting Modelfile overrides
is more complicated with the direct proxy approach)
- Updated CLI/TUI/config cloud detection to use shared selector logic
- Updated CLI preflight behavior so explicit cloud requests do not
auto-pull local stubs
What's next?
- Cloud discovery/listing and cache-backed `ollama ls` / `/api/tags`
- Modelfile overlay support for virtual cloud models on OpenAI/Anthropic
request families
- Recommender/default-selection behavior for ambiguous model families
- Fully remove the legacy flow
Fixes: https://github.com/ollama/ollama/issues/13801
* consolidate pull logic into confirmAndPull helper
pullIfNeeded and ShowOrPull shared identical confirm-and-pull logic.
Extract confirmAndPull to eliminate the duplication.
* skip local existence checks for cloud models
ModelExists and the TUI's modelExists both check the local model list,
which causes cloud models to appear missing. Return true early for
explicit cloud models so the TUI displays them beside the integration
name and skips re-prompting the model picker on relaunch.
* support optionally pulling stubs for newly-style names
We now normalize names like `<family>:<size>:cloud` into legacy-style
names like `<family>:<size>-cloud` for pulling and deleting (this also
supports stripping `:local`). Support for pulling cloud models is
temporary, once we integrate properly into `/api/tags` we won't need
this anymore.
* Fix server alias syncing
* Update cmd/cmd.go
Co-authored-by: Parth Sareen <parth.sareen@ollama.com>
* address comments
* improve some naming
---------
Co-authored-by: ParthSareen <parth.sareen@ollama.com>
The "(local)" qualifier is unnecessary since there's only one Ollama
provider. Existing configs with the old name are migrated automatically;
custom names are left unchanged.
Only the last token's processing time is included in prompt processing,
giving an artificially high rate. In addition, the number of tokens
only included the tokens that miss the cache, instead of our historic
total tokens.
Currently, context length is unbounded - the cache will keep
growing forever independent of the model's trained context
length. This caps it and enforces semantics similar to most
cloud services:
- Long prompts will result in an error, not truncation.
- Generation that exceeds the context will be stopped
Errors that occur during pipeline processing are currently only
logged but not sent back to the client. Rather than using HTTP
status codes as we have historically done, this serializes errors
as messages to allow sending them at any time during the stream.
The MLX runner previously reported a static VRAM estimate that was
computed at load time and consisted only of the weights. This is
strictly less than the actual memory usage, as it does not include
the KV cache or compute graph.
When the entire prompt was already cached (e.g. repeated prompt),
findRemaining returned an empty slice, causing FromValues to panic
on an index-out-of-range accessing a zero-length byte slice.
Fix by always keeping at least one token to re-evaluate so the
pipeline can seed token generation. Also reject empty prompts
early rather than panicking.
Align Qwen parser behavior with Transformers serve by allowing <tool_call> parsing while still in thinking collection.
Changes:
- qwen3vl: detect <tool_call> before </think> in thinking state and transition to tool parsing
- qwen3: same thinking-state tool detection and partial-tag overlap handling
- tests: update qwen3vl thinking/tool interleaving expectations
- tests: add qwen3 cases for tool call before </think> and split <tool_call> streaming
Currently, a canceled request can result in computation continuing
in the background to completion. It can also trigger a deadlock
when there is nobody to read the output tokens and the pipeline
cannot continue to the next request.
Particularly in error cases, it can be difficult to ensure that
all pinned memory is unpinned, MLX buffers are released and cache
state is consistent. This encapsulates those pieces and sets up
proper deferrals so that this happens automatically on exit.
Pass subprocess stdout/stderr through to the parent's stderr directly
instead of re-wrapping each line with slog. The subprocess already
writes structured slog output, so the re-wrapping produced nested
timestamps, levels, and message fields that were hard to read.
Also downgrade verbose KV cache debug logs to trace level.
The KV cache previously used a tree structure which could
store multiple divergent sequences, which is good for cache
reuse. However, this is typically used in conjunction with
paged attention so each node in the tree can store just a
chunk of the KV cache and they can be stitched together later.
We don't currently do this, so the cache was storing copies of
the full cache for each past sequence.
This redundancy plus the lack of resource limits, caused significant
memory use as a conversation grew. Instead, this changes to store
a single entry for the cache, which can be prefix matched. Although
it is less ideal for multiple users, it largely matches Ollama's
current behavior. It can be improved as additional pieces are fleshed
out.
The previous approach tracked array lifecycles through reference
counting, where each array recorded its inputs and a reference count
that was decremented as dependents were freed. This is not really
necessary as MLX tracks references internally. It is also error
prone as it is easy to create new arrays and forget to free them
when the Go variable goes out of scope.
Instead, we can pin just the arrays we want (typically outputs and
specific intermediates, like the cache). All other arrays are freed
by default when we run sweep. This avoids most causes of memory leaks
while still giving the freedom to save what we want.
The recent change in #14322 added tryLoadByName() which attempts to
load libmlxc.dylib via rpath before searching directories. This is an
optimization for Homebrew installations where rpath is correctly set.
However, when rpath isn't set (which is the common case for app bundle
installations), dlopen fails and the CHECK macro prints an error to
stderr:
ERROR - dynamic.c:21 - CHECK failed: handle->ctx != NULL
This error is misleading because it's an expected failure path - the
code correctly falls back to searching the executable directory and
loads the library successfully. The error message causes user confusion
and makes it appear that something is broken.
Replace the CHECK macro with a simple return code so the C code fails
silently. The Go code already handles error logging appropriately:
tryLoadByName() fails silently (intentional fallback), while
tryLoadFromDir() logs via slog.Error() when explicit path loading fails.
Parse the default_num_ctx from the server's "vram-based default context"
log line and expose it through the inference compute API. This eliminates
duplicate VRAM tier calculation logic in the frontend.
- Add InferenceInfo struct with Computes and DefaultContextLength
- Rename GetInferenceComputer to GetInferenceInfo
- Handle missing default context line gracefully (older servers)
- Add DefaultContextLength to InferenceComputeResponse
- Update Settings UI to use server's default, disable slider while loading
- Add disabled prop to Slider component (grays out + hides handle)
- Migrate existing users with context_length=4096 to 0 (auto mode)
This change adds a new x/tokenizer package which includes:
* New BPE and SentencePiece tokenizers
* Removing the dependency on the imagegen tokenizers
* Fixes to multibyte decoding in the pipeline
* Various correctness and benchmark tests
Not included in this PR is the WordPiece tokenizer for BERT models which will be
added when we add embedding models. The imagegen tokenizers will also be removed in
a follow-up PR.
The existing code manually searches directories for libmlxc.* and passes
full paths to dlopen, bypassing the binary's rpath. This means MLX
libraries installed via package managers (e.g., Homebrew) aren't found
even when rpath is correctly set at link time.
This change adds a fallback that tries loading via rpath first (using
just the library name), before falling back to the existing directory
search. This follows standard Unix/macOS conventions and works with any
installation that sets rpath.
Fixes library loading on macOS with Homebrew-installed mlx-c without
requiring OLLAMA_LIBRARY_PATH environment variable.
Co-authored-by: Natl <nat@MacBook-Pro.local>
The Codex runner was not setting OPENAI_BASE_URL or OPENAI_API_KEY, this prevents Codex from sending requests to api.openai.com instead of the local Ollama server. This mirrors the approach used by the Claude runner.
Codex v0.98.0 sends zstd-compressed request bodies to the /v1/responses endpoint. Add decompression support in ResponsesMiddleware with an 8MB max decompressed size limit to prevent resource exhaustion.
