builders · review · first seen Oct 29, 09:32 UTC

Ollama adds LFM3 support on Linux and Windows

Ollama v0.40.13 shipped LFM3 support on Linux and Windows, including text-to-speech in lfm3-audio and lfm3-omni. If your work depends on one of those operating systems, the release gives you a reason to test an LFM3 workflow that the cited release note now says is supported.

L0 · primary1 source2 claims checked3 min read

What shipped

The v0.40.13 release note says Liquid AI’s LFM3 models now run in Ollama on Linux and Windows. It describes this as an addition to Apple silicon support, not as a replacement for it. The concrete change for a Linux or Windows builder is the stated ability to run LFM3 models in Ollama. The release excerpt does not narrow that statement to a particular Linux distribution or Windows configuration. [1]

The same release note specifically includes text-to-speech in lfm3-audio and lfm3-omni. That detail matters because it identifies a capability to test, rather than leaving builders to infer audio support from the general LFM3 announcement. It says those text-to-speech paths now run on Linux and Windows. It does not describe the inputs, outputs, or steps for using either model, so the shipped claim is support for the capability, not a documented end-to-end procedure in this excerpt. [1]

Why it matters if you build

If you are building on Linux or Windows and had set aside an LFM3 evaluation because of operating-system support, v0.40.13 is a useful point to revisit it. The practical next decision is whether your own LFM3 task runs in your Ollama setup, not whether the release note alone proves that it fits your application. The source establishes that LFM3 now runs on those systems; it does not establish suitability for a particular workload. Treat a successful run of your task as a separate finding. [1]

For a text-to-speech workflow, the named options are lfm3-audio and lfm3-omni. A builder can now put either one on a Linux or Windows test list without assuming that the general model-support statement covers speech by implication. Budget the evaluation around what the note leaves open: it supplies no performance figures, resource requirements, prices, or output-quality measurements. Those are questions to check against your own constraints before changing a working path. The release claim is that text-to-speech runs, not that it meets any particular cost or quality target. [1]

How to try it today

1. Use Ollama v0.40.13 for the evaluation; that is the version named by the release note.
2. On a Linux or Windows system, try an LFM3 model in Ollama and record whether your task runs.
3. If you need text-to-speech, test lfm3-audio on that system and keep its result separate from the general LFM3 test.
4. Test lfm3-omni text-to-speech separately if it is an option you are considering.
5. Compare the observed behavior with your own requirements before deciding to use either text-to-speech path.

What we checked

Every claim below was walked down the source ladder. The rung is how close the source is to the thing itself; the tense is the source's own.

1. shipped L0 · primary

   Liquid AI’s LFM3 models now run in Ollama on Linux and Windows.

   per GitHub releases · ollama/ollama (L0)
2. shipped L0 · primary

   Text-to-speech in lfm3-audio and lfm3-omni now runs on Linux and Windows.

   per GitHub releases · ollama/ollama (L0)

What is not confirmed

The supplied release excerpt does not give setup commands, supported hardware specifications, Linux distribution details, or Windows version details. It also does not report latency, memory use, speech quality, reliability, or cost for either text-to-speech option. None of those points should be treated as confirmed by the support announcement. The note establishes that LFM3 models, including the named text-to-speech capabilities, now run on Linux and Windows; it does not establish how well they will serve a specific deployment. [1][2]

Who should act, who should wait

Act now if you build on Linux or Windows, already have an LFM3 use case to test, and can check the result against your own requirements. Include lfm3-audio or lfm3-omni in that test if text-to-speech is the relevant capability. Wait before making a deployment decision if it depends on performance, resource use, quality, or setup details that this excerpt does not provide. The release gives you a supported starting point for evaluation, not evidence that your particular workflow is ready to ship. [1][2]

Sources

• v0.40.13 — GitHub releases · ollama/ollama L0 · primary