ollama - 💡(How to fix) Fix MLX runner fails to load dynamic library on Ollama 0.20.5 (Apple Silicon, Homebrew install) [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
ollama/ollama#15479Fetched 2026-04-11 06:12:35
View on GitHub
Comments
0
Participants
1
Timeline
1
Reactions
1
Participants
Timeline (top)
labeled ×1

Error Message

After upgrading from 0.20.3 to 0.20.5, MLX-based models fail to run due to a dynamic library loading error. Non-MLX models continue to work normally. Restarting the Ollama service does not resolve the issue. Error: 500 Internal Server Error: mlx runner failed: Error: MLX not available: failed to load MLX dynamic library (searched: [/opt/homebrew/Cellar/ollama/0.20.5/bin/lib/ollama /opt/homebrew/Cellar/ollama/0.20.5/bin /opt/homebrew/var/build/lib/ollama]) (exit: exit status 1) Error suggests missing or incorrectly packaged MLX dynamic library rather than a model issue

Fix Action

Fix / Workaround

Install or upgrade to Ollama 0.20.5 via Homebrew Start the Ollama service (launchctl or equivalent) Run: ollama run qwen3.5:35b-a3b-mlx-bf16 "Write one sentence about testing."

RAW_BUFFERClick to expand / collapse

What is the issue?

Environment:

Hardware: Apple Silicon (M4 Max) OS: macOS (latest) Install method: Homebrew (official formula) Ollama version: 0.20.5 Previous working version: 0.20.3 (local Homebrew formula build)

Summary: After upgrading from 0.20.3 to 0.20.5, MLX-based models fail to run due to a dynamic library loading error. Non-MLX models continue to work normally. Restarting the Ollama service does not resolve the issue.

Reproduction steps:

Install or upgrade to Ollama 0.20.5 via Homebrew Start the Ollama service (launchctl or equivalent) Run: ollama run qwen3.5:35b-a3b-mlx-bf16 "Write one sentence about testing."

Expected behavior: The MLX BF16 model should load and run inference normally.

Actual behavior: Error: 500 Internal Server Error: mlx runner failed: Error: MLX not available: failed to load MLX dynamic library (searched: [/opt/homebrew/Cellar/ollama/0.20.5/bin/lib/ollama /opt/homebrew/Cellar/ollama/0.20.5/bin /opt/homebrew/var/build/lib/ollama]) (exit: exit status 1)

Additional observations:

Ollama server starts successfully and listens on 127.0.0.1:11434 Standard (non-MLX) models such as qwen3.5:35b run without issue Failure is specific to MLX-backed models Behavior persists after a full service restart Error suggests missing or incorrectly packaged MLX dynamic library rather than a model issue

Working comparison (0.20.3):

Same MLX BF16 model runs correctly on 0.20.3 No configuration changes between versions

Impact:

Breaks MLX-based workflows on Apple Silicon Prevents use of BF16 models Blocks upgrade from 0.20.3 for MLX users

Possible cause: This appears to be a packaging or runtime path issue affecting MLX dynamic library loading in the 0.20.5 Homebrew build.

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

extent analysis

TL;DR

Downgrade to Ollama version 0.20.3 to restore functionality of MLX-based models until the dynamic library loading issue in version 0.20.5 is resolved.

Guidance

  • Verify that the issue is specific to MLX-backed models by attempting to run a non-MLX model, as the problem seems to be isolated to MLX models.
  • Check the installation paths and library dependencies for the Ollama 0.20.5 version to ensure that the MLX dynamic library is correctly packaged and accessible.
  • Consider reinstalling Ollama 0.20.5 via Homebrew to ensure a clean installation, although the issue suggests a packaging problem rather than an installation error.
  • If downgrading to 0.20.3 is not feasible due to other dependencies, monitor the Ollama project for updates that address the MLX dynamic library loading issue in version 0.20.5.

Example

No specific code snippet is applicable in this scenario as the issue appears to be related to packaging and library loading rather than code implementation.

Notes

The solution provided is based on the information given and the observed behavior. The actual fix may depend on further investigation into the packaging and distribution of the MLX dynamic library with the Ollama 0.20.5 Homebrew build.

Recommendation

Apply workaround: Downgrade to version 0.20.3, as it is confirmed to work with MLX-based models, allowing for continued use of these models until the issue in version 0.20.5 is resolved.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

ollama - 💡(How to fix) Fix MLX runner fails to load dynamic library on Ollama 0.20.5 (Apple Silicon, Homebrew install) [1 participants]