Core ML
Shares rank 17 of 19 machine learning libraries by recommenders
People recommend it for running many models on Apple hardware, including the Neural Engine, but critics say it gives little control over which chip runs each layer.
People recommend it for running many models on Apple hardware, including the Neural Engine, but critics say it gives little control over which chip runs each layer.
15 people gave an opinion
- recommend it
- 5
- criticize it
- 10
- approval share
- 33%
Recommended in 5 of 13 months
- November 2025: not recommended
- December 2025: not recommended
- January 2026: not recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: not recommended
- June 2026: recommended
- July 2026: not recommended
- August 2026: not recommended
- September 2026: not recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 5 people recommend itNeeds 10 or more, not met.
- 0.5 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 5 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Runs models on Neural Engine
“I would most likely bet on CoreML since it have support for ANE”
Comment from 10 Mar 2026 in “Show HN: RunAnwhere – Faster AI Inference on Apple Silicon” “converting them to CoreML so they run in the Neural Engine”
Comment from 13 Oct 2026 in “We turned EmbeddingGemma 2 into a speech-to-text model” Handles many model types
“It is very suitable and very straightforward for a crazy large amount of models!”
Comment from 14 Oct 2026 in “We turned EmbeddingGemma 2 into a speech-to-text model” “it works for that without much trouble”
Comment from 9 Jun 2026 in “Apple Core AI Framework”
What people warn about
Each mark is one comment
Little control over hardware
“CoreML didn't give you much control (or even feedback) on what parts of your model would run on the CPU, GPU, or ANE”
Comment from 13 Sep 2026 in “Retrospectively Reverse-Engineering Apple's Neural Engine” “CoreML treats it as a black-box scheduler, stripping away any direct control or ability to train.”
Comment from 5 Mar 2026 in “Show HN: Orion – Native Training LLMs on the Apple Neural Engine Without CoreML” Bugs and crashes
“tracking down an uncatchable C++ assertion in Apple's CoreML stack”
Comment from 29 Mar 2026 in “An uncatchable CoreML crash: MLIR compiler failures on the iPhone SE 2” “Mac's CoreML has had soo many bugs over the last couple years”
Comment from 11 Aug 2026 in “Mojo 1.0 Is Here” Weak for most use cases
“ANE and CoreML is kind of meh for most use cases”
Comment from 4 Mar 2026 in “Apple introduces the new MacBook Air with M5” “Also generally I think CoreML isn't the best.”
Comment from 22 Dec 2025 in “ONNX Runtime and CoreML May Silently Convert Your Model to FP16”
Other machine learning libraries people recommend
The four most recommended other machine learning libraries, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
See all 19 machine learning libraries, rankedWhat people said
19 of the 20 comments that mention it, newest first. Each links to the original comment.
“It is very suitable and very straightforward for a crazy large amount of models!”
“CoreML is not suitable for all layers, so "converting them" is a) not straightforward; b) does not benefit the full set of layers.”
“converting them to CoreML so they run in the Neural Engine”
“Google's investment in Tensorflow Lite has paid off a lot better than MLX or CoreML did.”
“CoreML didn't give you much control (or even feedback) on what parts of your model would run on the CPU, GPU, or ANE”
“Mac's CoreML has had soo many bugs over the last couple years”
“even 10 years after CoreML was reluctantly introduced”
“libraries like MLX and CoreML built precariously overtop”
“I got some massive performance improvements as well.”
“Been doing that for years with CoreML”
“it works for that without much trouble”
“coreml are secure but they can only do the actual model not all the code around it”
“tracking down an uncatchable C++ assertion in Apple's CoreML stack”
“CoreML is optimized for classification and vision models, not autoregressive text generation”
“I would most likely bet on CoreML since it have support for ANE”
“CoreML treats it as a black-box scheduler, stripping away any direct control or ability to train.”
“ANE and CoreML is kind of meh for most use cases”
“Apple basically limits the runtime usage to ensure that no third party apps inhibit or restrict Apple's own use of this hardware.”
“Also generally I think CoreML isn't the best.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.