brew Ollama is missing llama-server

[thodem]

brew config AND brew doctor output OR brew gist-logs <formula> link

-- ~ » brew config
HOMEBREW_VERSION: 5.1.14
ORIGIN: https://github.com/Homebrew/brew
HEAD: 10a163ac127624caa80cc5cc5a705e97f3615b0e
Last commit: 9 days ago
Branch: stable
Core tap JSON: 02 Jun 08:42 UTC
Core cask tap JSON: 02 Jun 08:42 UTC
HOMEBREW_PREFIX: /opt/homebrew
HOMEBREW_CASK_OPTS: []
HOMEBREW_DOWNLOAD_CONCURRENCY: 32
HOMEBREW_FORBID_PACKAGES_FROM_PATHS: set
HOMEBREW_MAKE_JOBS: 16
Homebrew Ruby: 4.0.5 => /opt/homebrew/Library/Homebrew/vendor/portable-ruby/4.0.5_1/bin/ruby
CPU: 16-core 64-bit arm_brava
Clang: 21.0.0 build 2100
Git: 2.54.0 => /opt/homebrew/bin/git
Curl: 8.7.1 => /usr/bin/curl
macOS: 26.5-arm64
CLT: 26.5.0.0.1777544298
Xcode: 26.5
Metal Toolchain: N/A
Rosetta 2: false

Verification

• My brew doctor output says Your system is ready to brew. and am still able to reproduce my issue.
• I ran brew update and am still able to reproduce my issue.
• I have resolved all warnings from brew doctor and that did not fix my problem.
• I searched for recent similar issues at https://github.com/Homebrew/homebrew-core/issues?q=is%3Aissue and found no duplicates.
• My issue is not about a failure to build a formula from source.

What were you trying to do (and why)?

since ollama 0.30, GGUF models fail to work because ollama needs llama-server bundled which wasn't needed before.
MLX models still run.

• see GGUF models doesnt'work anymore ollama/ollama#16417 (comment)

ollama run gemma4:26b-a4b-it-q8_0
leads to a failure ~"error starting llama-server: llama-server binary not found"

while
ollama run Gemma4:26b-mlx
works.

ollama needs a special llama-server, see ollama/ollama#16432 (comment)

What happened (include all command output)?

ollama run gemma4:26b-a4b-it-q8_0:

Error: 500 Internal Server Error: error starting llama-server: llama-server binary not found (checked: /opt/homebrew/Cellar/ollama/0.30.0/libexec/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/libexec/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/libexec/build/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/libexec/dist/darwin-arm64/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/libexec/dist/darwin_arm64/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.0/libexec/dist/darwin/llama-server, /opt/homebrew/var/build/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin-arm64/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin_arm64/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin/llama-server). Run 'cmake -S llama/server --preset cpu && cmake --build --preset cpu' first

What did you expect to happen?

promt is shown:

send a message

Step-by-step reproduction instructions (by running brew commands)

brew install ollama
ollama pull gemma4:26b-a4b-it-q8_0
ollama run gemma4:26b-a4b-it-q8_0

[eaglstun]

Can confirm this on Apple Silicon. Since 0.30, GGUF models need a separate llama-server binary (MLX models don't, which is why those still work). The formula's go build produces the MLX runner but not llama-server — Ollama builds that separately via the CMake project in llama/server/, so it never ends up in libexec/lib/ollama/.

Worth noting it has to be Ollama's own llama-server build rather than a stock llama.cpp — they patch it and pin it to a specific GGML version. Upstream context: ollama/ollama#16417 and ollama/ollama#16432.

[thodem]

in Ollama 0.30.5 it is still present.

 ollama run mistral-small3.2:24b
Error: 500 Internal Server Error: error starting llama-server: llama-server binary not found (checked: /opt/homebrew/Cellar/ollama/0.30.5/libexec/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/libexec/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/libexec/build/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/libexec/dist/darwin-arm64/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/libexec/dist/darwin_arm64/lib/ollama/llama-server, /opt/homebrew/Cellar/ollama/0.30.5/libexec/dist/darwin/llama-server, /opt/homebrew/var/build/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin-arm64/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin_arm64/lib/ollama/llama-server, /opt/homebrew/var/dist/darwin/llama-server). Run 'cmake -S llama/server --preset cpu && cmake --build --preset cpu' first
--- ~ ollama -v                                                                                                                                                                                        1 ↵
ollama version is 0.30.5

Mac OS Tahoe 26.5.1 / M4 Max

[hcguersoy]

My solution: fall back to version 0.24.0:

• Create a local tap

brew tap-new $USER/local-tap

• Enable local core tap - something about ~ 1.5 Gb !

brew tap --force homebrew/core

• Extract the specific Ollama formulae version for v0.24.0

HOMEBREW_NO_INSTALL_FROM_API=1 brew extract --version=0.24.0 ollama $USER/local-tap

• Install

brew install $USER/local-tap/ollama@0.24.0

• Pin the version

brew pin $USER/local-tap/ollama@0.24.0

• clean up (if you don't need the homebrew/core clone anymore)

brew untap homebrew/core
brew cleanup

This is simply build-in functionality in Homebrew and can be used for every formulae you need to fallback to an older formulae version.

[senki]

You don’t need to enable local core tap.

1. Create local tap as described by @hcguersoy.
2. Download the old rb file from GitHub ollama.rb 0.24.0.
3. Move it to your tap directory.
4. Install it with brew install $USER/local-tap/ollama.
5. You don’t need to pin it either.

To revert to the official version, simply remove it from brew and install it normally.

Edit to add: Your new tap directory is $(brew --taps)/$USER/local-tap/Formula/

[SMillerDev]

So, this in a nutshell with brew supported commands:

brew tap-new user/local-tap
brew extract --version=0.24.0 ollama user/local-tap
brew install user/local-tap/ollama@0.24.0

[hcguersoy]

So, this in a nutshell with brew supported commands:

brew tap-new user/local-tap
brew extract --version=0.24.0 ollama user/local-tap
brew install user/local-tap/ollama@0.24.0

extract only works if you've a local clone of the tap homebrew/core - which you get with brew tap --force homebrew/core.
The alternative is to do it without extract as described by @senki.

[Nemesis19]

ollama 0.30.6 still has the bug

[mikewalks]

ollama 0.30.6 still has the bug

You can follow the progress of #285963 for a fix, the releases of ollama so far are independent of that. Last update was that there are some problems with it failing a check for MacOS14 and that some help is requested. Devs also likely (hopefully) enjoying their weekend.

[ogiboy]

just install it from ollama itself. I just did it and it works like as it was before.