Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Qwen 2.5 Coder Q4 M4 Max Inference test. Apple MLX vs Ollama: - MLX: 23.97 toks/sec 🥇🔥 - Ollama: 18.33 toks/sec 🥈 Here a video to show results

Ivan Fioravanti ᯅ

15,122 subscribers

34,881 views • 1 year ago •via X (Twitter)

Education Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Qwen 32B (4-bit) generates at >40 toks/sec on an M4 Max with assisted decoding and Qwen 0.5B as the draft model. Coming soon to mlx-lm. Compare regular decoding (left) to assisted decoding (right):

Qwen 32B (4-bit) generates at >40 toks/sec on an M4 Max with assisted decoding and Qwen 0.5B as the draft model. Coming soon to mlx-lm. Compare regular decoding (left) to assisted decoding (right):

Awni Hannun

50,353 views • 1 year ago

Llama 3.3 70B 4-bit runs nicely on a 64GB M3 Max with in MLX LM (~10 toks/sec). Would be even faster on an M4 Max. Yesterday's server-only 405B is today's laptop 70B:

Llama 3.3 70B 4-bit runs nicely on a 64GB M3 Max with in MLX LM (~10 toks/sec). Would be even faster on an M4 Max. Yesterday's server-only 405B is today's laptop 70B:

Awni Hannun

110,375 views • 1 year ago

Llama 3.3 70B 4-bit running on a 128GB M4 Max with MLX LM (~12 toks/sec) with mactop on the right. "Good enough", local, private and open models are now more accessible than ever!

Llama 3.3 70B 4-bit running on a 128GB M4 Max with MLX LM (~12 toks/sec) with mactop on the right. "Good enough", local, private and open models are now more accessible than ever!

Ivan Fioravanti ᯅ

43,929 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:

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

196,482 views • 10 months ago

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!

Awni Hannun

168,842 views • 1 year ago

oMLX last version seems having some issues in decoding. 🧐 Here a test with Qwen3.6-35B-A3B-MLX-6bit on M5 Max using. 🥇 LM Studio MLX 1.8.5 → 100.9 toks/s 🥈 mlx-vlm 0.6.2 → 100.1 toks/s 🥉 oMLX 0.4.2 dev3 → 58.7 toks/s 👀 Avg Gen TPS: oMLX 58.7 → LM Studio 100.9 (+71.8%) I have to thank pymike00 that raised this oMLX issue after seeing my video on using it with Codex. I bet there is a bug in oMLX chat and server at the moment, because internal benchmarks are ok, video attached. I bet Jun Kim will fix this soon 💪

oMLX last version seems having some issues in decoding. 🧐 Here a test with Qwen3.6-35B-A3B-MLX-6bit on M5 Max using. 🥇 LM Studio MLX 1.8.5 → 100.9 toks/s 🥈 mlx-vlm 0.6.2 → 100.1 toks/s 🥉 oMLX 0.4.2 dev3 → 58.7 toks/s 👀 Avg Gen TPS: oMLX 58.7 → LM Studio 100.9 (+71.8%) I have to thank pymike00 that raised this oMLX issue after seeing my video on using it with Codex. I bet there is a bug in oMLX chat and server at the moment, because internal benchmarks are ok, video attached. I bet Jun Kim will fix this soon 💪

Ivan Fioravanti ᯅ

19,756 views • 17 days ago

M4 Mac AI Coding Cluster Uses EXO Labs to run LLMs (here Qwen 2.5 Coder 32B at 18 tok/sec) distributed across 4 M4 Mac Minis (Thunderbolt 5 80Gbps) and a MacBook Pro M4 Max. Local alternative to Cursor (benchmark comparison soon).

M4 Mac AI Coding Cluster Uses EXO Labs to run LLMs (here Qwen 2.5 Coder 32B at 18 tok/sec) distributed across 4 M4 Mac Minis (Thunderbolt 5 80Gbps) and a MacBook Pro M4 Max. Local alternative to Cursor (benchmark comparison soon).

Alex Cheema

517,325 views • 1 year ago

DeepSeek R1 (the 680B MOE) is ~20% faster in the latest mlx / mlx-lm. 4-bit model on 3 M2 Ultras generates 4k tokens at a respectable 15 toks/sec. Plus some QoL improvements: - Only downloads the local shard (much faster startup) - Distributed launcher ships with MLX

DeepSeek R1 (the 680B MOE) is ~20% faster in the latest mlx / mlx-lm. 4-bit model on 3 M2 Ultras generates 4k tokens at a respectable 15 toks/sec. Plus some QoL improvements: - Only downloads the local shard (much faster startup) - Distributed launcher ships with MLX

Awni Hannun

86,921 views • 1 year ago

QLoRA fine-tuning 4-bit Gemma 2B on iPhone 15 Pro with MLX Swift. A nice size for fine-tuning on device, getting 70-100 toks/sec depending on the batch. Guide here:

QLoRA fine-tuning 4-bit Gemma 2B on iPhone 15 Pro with MLX Swift. A nice size for fine-tuning on device, getting 70-100 toks/sec depending on the batch. Guide here:

Awni Hannun

19,530 views • 2 years 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

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

What is this Nanbeige4.1-3B model running at - 77 toks/s in bf16 (in video) - 115 toks/s in 8bit on M3 Ultra with MLX with these benchmark scores! 🔥

What is this Nanbeige4.1-3B model running at - 77 toks/s in bf16 (in video) - 115 toks/s in 8bit on M3 Ultra with MLX with these benchmark scores! 🔥

Ivan Fioravanti ᯅ

85,228 views • 4 months ago

Reasoning on Apple Vision Pro with Apple MLX and DeepSeek R1 Qwen 7B 4bit! 🔥 14 tokens per sec! 🔥 Note: sorry for the shaking footage, I was excited to see it running 😂

Reasoning on Apple Vision Pro with Apple MLX and DeepSeek R1 Qwen 7B 4bit! 🔥 14 tokens per sec! 🔥 Note: sorry for the shaking footage, I was excited to see it running 😂

Ivan Fioravanti ᯅ

24,324 views • 1 year ago

Running Ring 1T reasoning model on a single M3 Ultra with mlx-lm. It's quantized to 3.5 bits-per-weight. Uses 440GB and generated ~6k tokens at 18.2 toks/sec. Getting closer to GPT-5 at home.

Running Ring 1T reasoning model on a single M3 Ultra with mlx-lm. It's quantized to 3.5 bits-per-weight. Uses 440GB and generated ~6k tokens at 18.2 toks/sec. Getting closer to GPT-5 at home.

Awni Hannun

55,131 views • 8 months ago

MLX Swift example can also QLoRA fine-tune Llama 3.2. Here's the 1B fine-tuning on my iPhone 15 Pro at > 150 toks/sec. A this rate only takes a few minutes to learn some decent adapters fully on-device.

MLX Swift example can also QLoRA fine-tune Llama 3.2. Here's the 1B fine-tuning on my iPhone 15 Pro at > 150 toks/sec. A this rate only takes a few minutes to learn some decent adapters fully on-device.

Awni Hannun

91,632 views • 1 year ago

I bought the M4 Max 48GB RAM and first thing I did is to test local LLMs. It can barely run llama3.3 70b, but qwen2.5-coder 32b works very well. I'll test with MLX next ✌️

I bought the M4 Max 48GB RAM and first thing I did is to test local LLMs. It can barely run llama3.3 70b, but qwen2.5-coder 32b works very well. I'll test with MLX next ✌️

Daniel Nguyen

315,827 views • 1 year ago

Pretty cool that with the new Qwen 2.5 models you can ask questions / generate using a reasonably sized code-base as context, all running on a laptop with mlx-lm. The 7B runs pretty fast on an M4 Max using the mlx-lm code base (~16k lines) as context:

Pretty cool that with the new Qwen 2.5 models you can ask questions / generate using a reasonably sized code-base as context, all running on a laptop with mlx-lm. The 7B runs pretty fast on an M4 Max using the mlx-lm code base (~16k lines) as context:

Awni Hannun

27,442 views • 1 year ago

🤯 Wow! In one prompt Qwen3-Coder-Next generated a fully working flappy birds game in HTML. (0:05) Claude Code with Qwen3-Coder-Next (0:26) Shows the game running Run it fully locally: ollama pull qwen3-coder-next Ollama's cloud if you can't run it locally: ollama pull qwen3-coder-next:cloud Try launching it with Claude Code using ollama launch (link to play 🧵) So cool! Qwen Tongyi Lab Junyang Lin

🤯 Wow! In one prompt Qwen3-Coder-Next generated a fully working flappy birds game in HTML. (0:05) Claude Code with Qwen3-Coder-Next (0:26) Shows the game running Run it fully locally: ollama pull qwen3-coder-next Ollama's cloud if you can't run it locally: ollama pull qwen3-coder-next:cloud Try launching it with Claude Code using ollama launch (link to play 🧵) So cool! Qwen Tongyi Lab Junyang Lin

ollama

117,339 views • 4 months ago

🔥 DeepSeek R1 671B Q4 - M3 Ultra 512GB with MLX🔥 - 18.43 tokens/sec - Generates a p5js zero-shot, tested at video's end 😱 - Video in real-time, no acceleration! - First test, and I'm blown away! Prompt: "Create an amazing animation using p5js"

🔥 DeepSeek R1 671B Q4 - M3 Ultra 512GB with MLX🔥 - 18.43 tokens/sec - Generates a p5js zero-shot, tested at video's end 😱 - Video in real-time, no acceleration! - First test, and I'm blown away! Prompt: "Create an amazing animation using p5js"

Ivan Fioravanti ᯅ

46,451 views • 1 year ago