Awni Hannun's banner

Awni Hannun

@awnihannun • 44,814 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 просмотров

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 просмотров

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 просмотров

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 просмотров

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 просмотров

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).

35,231 просмотров

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:

46,777 просмотров

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 просмотров

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 просмотров

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 просмотров

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 просмотров

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

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

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

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:

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:

997,284 просмотров • 1 год назад

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:

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:

500,973 просмотров • 10 месяцев назад

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)

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)

866,192 просмотров • 1 год назад

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.

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.

243,254 просмотров • 7 месяцев назад

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.

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.

237,191 просмотров • 8 месяцев назад

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)

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)

581,777 просмотров • 2 лет назад

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

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

215,597 просмотров • 11 месяцев назад

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.

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.

141,194 просмотров • 7 месяцев назад

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:

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:

238,187 просмотров • 1 год назад

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:

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:

196,508 просмотров • 1 год назад

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:

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:

186,772 просмотров • 1 год назад

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

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

95,092 просмотров • 8 месяцев назад

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:

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:

94,109 просмотров • 8 месяцев назад

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

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

168,896 просмотров • 1 год назад

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:

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:

149,855 просмотров • 1 год назад

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.

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.

60,599 просмотров • 6 месяцев назад

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:

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:

132,923 просмотров • 1 год назад

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.

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.

48,673 просмотров • 6 месяцев назад

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?

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?

68,539 просмотров • 11 месяцев назад