Awni Hannun's banner
Awni Hannun's profile picture

Awni Hannun

@awnihannun • 44,906 subscribers

ow knee

Shorts

DeepSeek R1 distilled to Qwen 1.5B easily runs on my iPhone 16 with MLX swift. Here's the 4-bit model reasoning entirely on device at almost 60 toks/sec:

DeepSeek R1 distilled to Qwen 1.5B easily runs on my iPhone 16 with MLX swift. Here's the 4-bit model reasoning entirely on device at almost 60 toks/sec:

1,139,656 views

Llama 3.2 1B in 4-bit runs at ~60 toks/sec with MLX Swift on my iPhone 15 pro. It's quite good and easily runs on-device:

Llama 3.2 1B in 4-bit runs at ~60 toks/sec with MLX Swift on my iPhone 15 pro. It's quite good and easily runs on-device:

492,428 views

mlx_lm.server + Qwen3 Coder Next 6bit + OpenCode + M3 Ultra = a pretty capable and very fast local coding setup. Not sped up:

mlx_lm.server + Qwen3 Coder Next 6bit + OpenCode + M3 Ultra = a pretty capable and very fast local coding setup. Not sped up:

66,668 views

Managed to get Ling Mini 16B (1.4B active) running on my iPhone Air. It runs very fast with MLX. It's a DWQ of Ling Mini quantized to 3 bits-per-weight. A 16B model running on an Air at this speed is pretty awesome:

Managed to get Ling Mini 16B (1.4B active) running on my iPhone Air. It runs very fast with MLX. It's a DWQ of Ling Mini quantized to 3 bits-per-weight. A 16B model running on an Air at this speed is pretty awesome:

92,422 views

We (Claude and I) made a new package: voxmlx It's an MLX implementation of Mistral's Voxtral mini realtime speech recognition model. It supports streaming audio and runs pretty fast on a laptop. To use it, simply: `uvx voxmlx` Also I did not write a single line of code for this package. Every line of code was written by Claude Code. More on that in thread.

We (Claude and I) made a new package: voxmlx It's an MLX implementation of Mistral's Voxtral mini realtime speech recognition model. It supports streaming audio and runs pretty fast on a laptop. To use it, simply: `uvx voxmlx` Also I did not write a single line of code for this package. Every line of code was written by Claude Code. More on that in thread.

33,896 views

Got continuous batching working with SSMs in mlx-lm. Here's four OpenCode agents simultaneously running Nvidia's Nemotron Nano on 64GB M4 Max. This is a nice model for smaller machines since it's MoE + hybrid attention (small cache).

Got continuous batching working with SSMs in mlx-lm. Here's four OpenCode agents simultaneously running Nvidia's Nemotron Nano on 64GB M4 Max. This is a nice model for smaller machines since it's MoE + hybrid attention (small cache).

35,078 views

No-one: But can you do 16 generations on your M4 laptop simultaneously? MLX LM:

No-one: But can you do 16 generations on your M4 laptop simultaneously? MLX LM:

46,767 views

Cool video (sped-up) running Mixtral 8x22B on a M3 Max (h/t PSA since I get this question a lot: asitop is the monitoring tool on the right (pip install asitop) glances is on the bottom (pip install glances)

Cool video (sped-up) running Mixtral 8x22B on a M3 Max (h/t PSA since I get this question a lot: asitop is the monitoring tool on the right (pip install asitop) glances is on the bottom (pip install glances)

70,804 views

Gotta love MoEs on Apple silicon with MLX. Kimi's new 16B (3B active) Moonshot model runs very nicely on an M4 Max. As good or better than some of the best dense 7Bs and 1.5x faster inference (154 toks/sec!):

Gotta love MoEs on Apple silicon with MLX. Kimi's new 16B (3B active) Moonshot model runs very nicely on an M4 Max. As good or better than some of the best dense 7Bs and 1.5x faster inference (154 toks/sec!):

23,209 views

MLX Swift LLM example works with: - Mistral / Llama - Phi-2 - Qwen 1.5 - Starcoder 2 Quick-start: Qwen 1.5 0.5B runs pretty fast in 16-bit on my iPhone 14, no quantization needed:

MLX Swift LLM example works with: - Mistral / Llama - Phi-2 - Qwen 1.5 - Starcoder 2 Quick-start: Qwen 1.5 0.5B runs pretty fast in 16-bit on my iPhone 14, no quantization needed:

30,441 views

DeepSeek-Prover (4-bit 7B) running at 114 toks/sec in MLX LM on an M2 Ultra Thanks to for the port!

DeepSeek-Prover (4-bit 7B) running at 114 toks/sec in MLX LM on an M2 Ultra Thanks to for the port!

16,079 views

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

The new 1 Trillion parameter Kimi K2 Thinking model runs well on 2 M3 Ultras in its native format - no loss in quality! The model was quantization aware trained (qat) at int4. Here it generated ~3500 tokens at 15 toks/sec using pipeline-parallelism in mlx-lm:
0:41
awnihannun's profile picture

The new 1 Trillion parameter Kimi K2 Thinking model runs well on 2 M3 Ultras in its native format - no loss in quality! The model was quantization aware trained (qat) at int4. Here it generated ~3500 tokens at 15 toks/sec using pipeline-parallelism in mlx-lm:

Awni Hannun

500,804 views • 8 months ago

DeepSeek R1 (the full 680B model) runs nicely in higher quality 4-bit on 3 M2 Ultras with MLX. Asked it a coding question and it thought for ~2k tokens and generated 3500 tokens overall:
1:28
awnihannun's profile picture

DeepSeek R1 (the full 680B model) runs nicely in higher quality 4-bit on 3 M2 Ultras with MLX. Asked it a coding question and it thought for ~2k tokens and generated 3500 tokens overall:

Awni Hannun

997,198 views • 1 year ago

DeepSeek R1 671B running on 2 M2 Ultras faster than reading speed. Getting close to open-source O1, at home, on consumer hardware. With mlx.distributed and mlx-lm, 3-bit quantization (~4 bpw)
0:54
awnihannun's profile picture

DeepSeek R1 671B running on 2 M2 Ultras faster than reading speed. Getting close to open-source O1, at home, on consumer hardware. With mlx.distributed and mlx-lm, 3-bit quantization (~4 bpw)

Awni Hannun

866,067 views • 1 year ago

Kimi K2.5 1T runs on 2 M3 Ultras with mlx-lm in it's native precision. It's actually quite usable. Here it's making a space invaders game. Generated 3856 tokens at 21.9 tok/sec using 350GB per machine. Thanks to Tarjei Mandt for the port.
1:08
awnihannun's profile picture

Kimi K2.5 1T runs on 2 M3 Ultras with mlx-lm in it's native precision. It's actually quite usable. Here it's making a space invaders game. Generated 3856 tokens at 21.9 tok/sec using 350GB per machine. Thanks to Tarjei Mandt for the port.

Awni Hannun

241,472 views • 5 months ago

GLM-4.7 runs quite well on an M3 Ultra with mlx-lm, even at a near lossless precision (6-bit here). It generated the best space invaders game I've seen yet for a local model (even included sound effects!). Generated 6600 tokens and ran at 16 tok/s.
0:53
awnihannun's profile picture

GLM-4.7 runs quite well on an M3 Ultra with mlx-lm, even at a near lossless precision (6-bit here). It generated the best space invaders game I've seen yet for a local model (even included sound effects!). Generated 6600 tokens and ran at 16 tok/s.

Awni Hannun

237,149 views • 6 months ago

Next level: QLoRA fine-tuning 4-bit Llama 3 8B on iPhone 15 pro. Incoming (Q)LoRA MLX Swift example by David Koski: works with lot's of models (Mistral, Gemma, Phi-2, etc)
1:19
awnihannun's profile picture

Next level: QLoRA fine-tuning 4-bit Llama 3 8B on iPhone 15 pro. Incoming (Q)LoRA MLX Swift example by David Koski: works with lot's of models (Mistral, Gemma, Phi-2, etc)

Awni Hannun

581,723 views • 2 years ago

GLM 4.7 Flash is supported in mlx-lm 0.30.3 (h/t Ivan Fioravanti ᯅ) The 4-bit runs fast (43 tok/s generation, ~800 tok/s prefill) on a base M5 32GB laptop.
0:37
awnihannun's profile picture

GLM 4.7 Flash is supported in mlx-lm 0.30.3 (h/t Ivan Fioravanti ᯅ) The 4-bit runs fast (43 tok/s generation, ~800 tok/s prefill) on a base M5 32GB laptop.

Awni Hannun

141,194 views • 6 months ago

Running Qwen3 8B thinking on an iPhone Air with MLX. The model is quantized to 4-bit and runs pretty well.
0:45
awnihannun's profile picture

Running Qwen3 8B thinking on an iPhone Air with MLX. The model is quantized to 4-bit and runs pretty well.

Awni Hannun

215,529 views • 10 months ago

The new Kimi K2 1T model (4-bit quant) runs on 2 512GB M3 Ultras with mlx-lm and mx.distributed. 1 trillion params, at a speed that's actually quite usable:
0:37
awnihannun's profile picture

The new Kimi K2 1T model (4-bit quant) runs on 2 512GB M3 Ultras with mlx-lm and mx.distributed. 1 trillion params, at a speed that's actually quite usable:

Awni Hannun

238,057 views • 1 year ago

Qwen3-Coder-Flash runs quite fast on an M4 Max with mlx-lm. Running the 4-bit here, generated 4,467 tokens at >107 tokens/sec:
0:21
awnihannun's profile picture

Qwen3-Coder-Flash runs quite fast on an M4 Max with mlx-lm. Running the 4-bit here, generated 4,467 tokens at >107 tokens/sec:

Awni Hannun

196,508 views • 11 months ago

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:
0:50
awnihannun's profile picture

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:

Awni Hannun

186,641 views • 1 year ago

Running four simultaneous OpenCode agents works well with mlx_lm.server continuous batching and MiniMax M2.1 on an M3 Ultra:
0:40
awnihannun's profile picture

Running four simultaneous OpenCode agents works well with mlx_lm.server continuous batching and MiniMax M2.1 on an M3 Ultra:

Awni Hannun

95,050 views • 6 months ago

MiniMax M2.1 in 4-bit cruises on an M3 Ultra with mlx-lm. Generated a space invaders game using 5098 tokens at 47.2 tok/sec:
0:23
awnihannun's profile picture

MiniMax M2.1 in 4-bit cruises on an M3 Ultra with mlx-lm. Generated a space invaders game using 5098 tokens at 47.2 tok/sec:

Awni Hannun

93,826 views • 6 months ago

The new Deep Seek V3 0324 in 4-bit runs at > 20 toks/sec on a 512GB M3 Ultra with mlx-lm!
0:31
awnihannun's profile picture

The new Deep Seek V3 0324 in 4-bit runs at > 20 toks/sec on a 512GB M3 Ultra with mlx-lm!

Awni Hannun

168,891 views • 1 year ago

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.
1:11
awnihannun's profile picture

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.

Awni Hannun

60,599 views • 5 months ago

First results are in. Llama 4 Maverick 17B active / 400B total is blazing fast with MLX on an M3 Ultra. Here is the 4-bit model generating 1100 tokens at 50 tok/sec:
0:20
awnihannun's profile picture

First results are in. Llama 4 Maverick 17B active / 400B total is blazing fast with MLX on an M3 Ultra. Here is the 4-bit model generating 1100 tokens at 50 tok/sec:

Awni Hannun

149,855 views • 1 year ago

QwQ-32B evals on par with Deep Seek R1 680B but runs fast on a laptop. Delivery accepted. Here it is running nicely on a M4 Max with MLX. A snippet of its 8k token long thought process:
0:53
awnihannun's profile picture

QwQ-32B evals on par with Deep Seek R1 680B but runs fast on a laptop. Delivery accepted. Here it is running nicely on a M4 Max with MLX. A snippet of its 8k token long thought process:

Awni Hannun

132,388 views • 1 year ago

Qwen3 235B MoE (22B active) runs so fast on an M2 Ultra with mlx-lm. - 4-bit model uses ~132GB - Generated 580 tokens at ~28 toks/sec
0:29
awnihannun's profile picture

Qwen3 235B MoE (22B active) runs so fast on an M2 Ultra with mlx-lm. - 4-bit model uses ~132GB - Generated 580 tokens at ~28 toks/sec

Awni Hannun

117,763 views • 1 year ago

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.
0:28
awnihannun's profile picture

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Awni Hannun

48,673 views • 5 months ago

GLM 4.6 runs quite fast on an M3 Ultra with mlx-lm even at higher precision. Pretty remarkable that it benchmarks competitive to the just-released Sonnet 4.5. Hope those benchmarks hold-up in day-to-day use. Here's a run using 5.5 bpw quantized model, generating 5.3k tokens at 17+ tok/sec using 244 GB. What prompts should I test?
0:54
awnihannun's profile picture

GLM 4.6 runs quite fast on an M3 Ultra with mlx-lm even at higher precision. Pretty remarkable that it benchmarks competitive to the just-released Sonnet 4.5. Hope those benchmarks hold-up in day-to-day use. Here's a run using 5.5 bpw quantized model, generating 5.3k tokens at 17+ tok/sec using 244 GB. What prompts should I test?

Awni Hannun

68,539 views • 9 months ago