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.

Just added Qwen 2.5 32B Coder to LlamaCoder – it's an amazing open source coding model. Going to run some evals between it and Llama 3.1 405B & will share my results soon.

Hassan

94,457 subscribers

52,565 views • 1 year ago •via X (Twitter)

Education News & Politics Science & Technology

Anya Rossi• Live Now

Private livecam show

10 Comments

Hassan1 year ago

You can try out Qwen 2.5 32B Coder on Together AI – we just brought it up yesterday!

Ryan Vogel1 year ago

I use Qwen locally with cursor and it is amazing!

Hassan1 year ago

Nice! Runs on your local machine?

fullstack1 year ago

So... in my testing, it works really good. But sometimes it switches or breaks variable names. I had a variable called 'context' and it changed it to 'text', and because it was multi threaded, I didn't catch the exception on the console. slightly annoying thing that claude wouldn't do.

Hassan1 year ago

Agreed! Sometimes you can tell it's a 32B model since it makes small mistakes (and is worse on reasoning). But still super impressive for how small it is

Tembo👨‍💻🚀1 year ago

@bytechicX

Ahmed Benhaliliba1 year ago

Check deepseek 2.5v too

race-of-sloths1 year ago

Impressive work, Nutlope! At Race of Sloth, we’re inspired by open-source projects that empower creators and foster innovation.

Mustafa1 year ago

qwen gonna probably win

Hassan1 year ago

@agent_mustafa maybe, there are benefits from a bigger model too especially w/ RAG on shadcn like I'm doing so I think it'll be close!

Related Videos

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

517,664 views • 1 year ago

🆕 How to run (and finetune) open source AI models with a simple API! In 5 mins, I go over how to: ◆ Generate text with DeepSeek R1 & Llama 3 ◆ Generate code with Qwen on LlamaCoder ◆ Generate images with Flux on BlinkShot ◆ Finetune a model on your own data & run it

🆕 How to run (and finetune) open source AI models with a simple API! In 5 mins, I go over how to: ◆ Generate text with DeepSeek R1 & Llama 3 ◆ Generate code with Qwen on LlamaCoder ◆ Generate images with Flux on BlinkShot ◆ Finetune a model on your own data & run it

Hassan

30,236 views • 1 year ago

Anthropic just dropped Claude Opus 4.1 It outperforms OpenAI o3, Gemini 2.5 Pro and Qwen-3 Coder on agentic coding and tool use. Claude Code is going to get incredibly better.

Anthropic just dropped Claude Opus 4.1 It outperforms OpenAI o3, Gemini 2.5 Pro and Qwen-3 Coder on agentic coding and tool use. Claude Code is going to get incredibly better.

Shubham Saboo

26,252 views • 1 year ago

100% Local LLM Coding with Bolt + Ollama + Qwen Qwen just dropped their qwen2.5-coder-30B and it's on ollama. Finally a model good enough to use instead of gpt-4o or claude. Now with Qwen, Bolt perfectly displays the app preview with Ollama. Run it with 1 click.

100% Local LLM Coding with Bolt + Ollama + Qwen Qwen just dropped their qwen2.5-coder-30B and it's on ollama. Finally a model good enough to use instead of gpt-4o or claude. Now with Qwen, Bolt perfectly displays the app preview with Ollama. Run it with 1 click.

cocktail peanut

52,630 views • 1 year ago

Announcing LlamaCoder v3 – generate React apps in 1 prompt! • Multi-file generation for better apps • Monaco editor to view & export code • New models: GLM 4.6, Kimi K2, & Qwen 3 Coder 100% free, open source, and powered by Together AI.

Announcing LlamaCoder v3 – generate React apps in 1 prompt! • Multi-file generation for better apps • Monaco editor to view & export code • New models: GLM 4.6, Kimi K2, & Qwen 3 Coder 100% free, open source, and powered by Together AI.

Hassan

42,450 views • 7 months ago

Introducing TogetherLink! An open source CLI to run any open source model inside your favorite coding harness. Run GLM 5.2 directly in Codex and Claude Code.

Introducing TogetherLink! An open source CLI to run any open source model inside your favorite coding harness. Run GLM 5.2 directly in Codex and Claude Code.

Hassan

33,312 views • 17 days ago

Alibaba just released a coding model that hits 82 percent on SWE-Bench Verified. That is the highest score ever published for an open-source model. The weights are free. The license is Apache 2.0. You can run it today. The model is Qwen 4 Coder 32B. Here is what 82 percent on SWE-Bench Verified actually means. SWE-Bench Verified tests whether an AI can autonomously resolve real bugs pulled from real production GitHub repositories. Not synthetic exercises. Real open-source projects that real teams depend on. A model gets a bug report, reads the code, writes a fix, and either passes the test suite or it does not. At 82 percent, Qwen 4 Coder 32B resolves 82 out of every 100 real production bugs it is given. Without a human guiding it. On code it has never seen before. For comparison: Qwen 4 Coder 32B: 82 percent SWE-Bench Verified. Open source. Apache 2.0. Claude Fable 5: 80.3 percent SWE-Bench Pro. $10 input / $50 output per million tokens. Currently suspended. GPT-5.6 Sol: Competitive on Terminal-Bench. $5 input / $30 output per million tokens. An open-weight model that you can download and run for free just beat both of them on the benchmark designed to measure real software engineering capability. Here is the architecture. Qwen 4 Coder 32B is a 32 billion parameter dense model. Not a Mixture-of-Experts. Every parameter is active on every request. This matters for inference: a dense 32B model runs on 22 gigabytes of VRAM, which fits on a single high-end consumer GPU or a MacBook Pro with 64GB of unified memory. The smaller variant, Qwen 4 Coder 4B, runs at approximately 135 tokens per second on an M5 Max and fits inside 8 gigabytes of RAM. For a model with usable coding capability, that is a new bar for what fits in a single laptop. The training methodology continued Alibaba's approach of reinforcement learning on verifiable coding tasks. The model gets rewarded when its code passes tests. It gets penalized when it fails. Over millions of training steps, the model learns to write code that actually runs rather than code that looks plausible. License: Apache 2.0. Full commercial use. No attribution requirement. No revenue threshold. No monthly active user ceiling. Weights: Hugging Face, available today. Runs on: vLLM, Ollama, SGLang, and any standard GGUF-compatible inference engine. Qwen 4 32B also runs at approximately 135 tokens per second on an M5 Max chip, setting a new bar for what a sub-8GB model can do on Apple Silicon. The open-source coding model just beat the best closed-source model in the world on the benchmark designed to test whether AI can actually do software engineering. The weights are free. The subscription is optional. Source: Autom8Labs AI Insight July 2026, State of Open Source LLMs June 2026, Kunal Ganglani blog June 2026.

Alibaba just released a coding model that hits 82 percent on SWE-Bench Verified. That is the highest score ever published for an open-source model. The weights are free. The license is Apache 2.0. You can run it today. The model is Qwen 4 Coder 32B. Here is what 82 percent on SWE-Bench Verified actually means. SWE-Bench Verified tests whether an AI can autonomously resolve real bugs pulled from real production GitHub repositories. Not synthetic exercises. Real open-source projects that real teams depend on. A model gets a bug report, reads the code, writes a fix, and either passes the test suite or it does not. At 82 percent, Qwen 4 Coder 32B resolves 82 out of every 100 real production bugs it is given. Without a human guiding it. On code it has never seen before. For comparison: Qwen 4 Coder 32B: 82 percent SWE-Bench Verified. Open source. Apache 2.0. Claude Fable 5: 80.3 percent SWE-Bench Pro. $10 input / $50 output per million tokens. Currently suspended. GPT-5.6 Sol: Competitive on Terminal-Bench. $5 input / $30 output per million tokens. An open-weight model that you can download and run for free just beat both of them on the benchmark designed to measure real software engineering capability. Here is the architecture. Qwen 4 Coder 32B is a 32 billion parameter dense model. Not a Mixture-of-Experts. Every parameter is active on every request. This matters for inference: a dense 32B model runs on 22 gigabytes of VRAM, which fits on a single high-end consumer GPU or a MacBook Pro with 64GB of unified memory. The smaller variant, Qwen 4 Coder 4B, runs at approximately 135 tokens per second on an M5 Max and fits inside 8 gigabytes of RAM. For a model with usable coding capability, that is a new bar for what fits in a single laptop. The training methodology continued Alibaba's approach of reinforcement learning on verifiable coding tasks. The model gets rewarded when its code passes tests. It gets penalized when it fails. Over millions of training steps, the model learns to write code that actually runs rather than code that looks plausible. License: Apache 2.0. Full commercial use. No attribution requirement. No revenue threshold. No monthly active user ceiling. Weights: Hugging Face, available today. Runs on: vLLM, Ollama, SGLang, and any standard GGUF-compatible inference engine. Qwen 4 32B also runs at approximately 135 tokens per second on an M5 Max chip, setting a new bar for what a sub-8GB model can do on Apple Silicon. The open-source coding model just beat the best closed-source model in the world on the benchmark designed to test whether AI can actually do software engineering. The weights are free. The subscription is optional. Source: Autom8Labs AI Insight July 2026, State of Open Source LLMs June 2026, Kunal Ganglani blog June 2026.

Harman

41,179 views • 24 days ago

You can now try Llama 3.1 405B for free (link below)! This is the largest open-source model out there, and for the first time, an open model is competitive with closed models. This time around, Meta did something new: Llama 3.1 has a license that allows developers to use it to enhance other models. For the first time, you can distill Llama 3.1 405B's capabilities into a smaller, more practical model for your use case. First, here is the link where you can play with Llama 3.1 for free: The model is hosted in Tune Studio, an end-to-end platform for developing applications using Large Language Models. They are sponsoring this post. Take a look at the attached video. It will show you how you can fine-tune a simple model using Llama 3.1 without leaving the platform: 1. You can create an empty dataset 2. Use the playground to generate and record interactions with Llama 3.1 3. Modify the dataset directly using the playground 4. Export the data and fine-tune a smaller model Fast and easy! As long as you have a web browser, you can start experimenting with fine-tuning and Llama 3.1. That's all it takes!

You can now try Llama 3.1 405B for free (link below)! This is the largest open-source model out there, and for the first time, an open model is competitive with closed models. This time around, Meta did something new: Llama 3.1 has a license that allows developers to use it to enhance other models. For the first time, you can distill Llama 3.1 405B's capabilities into a smaller, more practical model for your use case. First, here is the link where you can play with Llama 3.1 for free: The model is hosted in Tune Studio, an end-to-end platform for developing applications using Large Language Models. They are sponsoring this post. Take a look at the attached video. It will show you how you can fine-tune a simple model using Llama 3.1 without leaving the platform: 1. You can create an empty dataset 2. Use the playground to generate and record interactions with Llama 3.1 3. Modify the dataset directly using the playground 4. Export the data and fine-tune a smaller model Fast and easy! As long as you have a web browser, you can start experimenting with fine-tuning and Llama 3.1. That's all it takes!

Santiago

55,609 views • 2 years ago

Qwen just published the 'Thinking' variant of this model 🔥 So you can run a model EVEN MORE powerful than GPT-4o locally! - Still only 3B active parameters - Open source license - 256k context window extendable to 1M - Strong in math, science and coding Details below ↓

Qwen just published the 'Thinking' variant of this model 🔥 So you can run a model EVEN MORE powerful than GPT-4o locally! - Still only 3B active parameters - Open source license - 256k context window extendable to 1M - Strong in math, science and coding Details below ↓

Paul Couvert

106,261 views • 1 year ago

I used GLM-5.2 to optimize this visualization I previously made, and it doesn't feel like an open-source model at all. It's amazing.

I used GLM-5.2 to optimize this visualization I previously made, and it doesn't feel like an open-source model at all. It's amazing.

Joshua Guo

116,696 views • 1 month ago

running a 405-billion-parameter model used to mean a $30,000 GPU rack. one developer made it run on a single 8GB card the kind that's been sitting in gaming laptops for years. it's called AirLLM, and the trick is almost stupidly simple: instead of loading the whole model, it keeps one layer on the GPU at a time and streams the rest through. / Llama 3.1 405B on 8GB / DeepSeek-V3, all 671B of it, on about 12GB / no quantization, no distillation, no pruning the full-precision model / works with Llama, Qwen, DeepSeek, Mistral, Gemma, and more it's Apache 2.0 and already past 22,600 stars. → bookmark it.

running a 405-billion-parameter model used to mean a $30,000 GPU rack. one developer made it run on a single 8GB card the kind that's been sitting in gaming laptops for years. it's called AirLLM, and the trick is almost stupidly simple: instead of loading the whole model, it keeps one layer on the GPU at a time and streams the rest through. / Llama 3.1 405B on 8GB / DeepSeek-V3, all 671B of it, on about 12GB / no quantization, no distillation, no pruning the full-precision model / works with Llama, Qwen, DeepSeek, Mistral, Gemma, and more it's Apache 2.0 and already past 22,600 stars. → bookmark it.

dunik

148,663 views • 17 days ago

This is wild! Llama 3.1 405B Instruct finally solves a famous math puzzle that was originally posted on /LocalLlama. To the best of my knowledge, every model (including Claude 3.5 Sonnet and GPT-4o) fails at this task. A longer video coming soon!

This is wild! Llama 3.1 405B Instruct finally solves a famous math puzzle that was originally posted on /LocalLlama. To the best of my knowledge, every model (including Claude 3.5 Sonnet and GPT-4o) fails at this task. A longer video coming soon!

elvis

52,546 views • 2 years ago

Hunyuan-Large by Tencent is a 389B param MOE (52B active). It's the largest open-weights MOE. In some benchmarks it exceeds Llama 3.1 405B. With MLX's new 3-bit quant it just barely fits on a single 192GB M2 Ultra! And runs at a very decent >15 toks/sec:

Hunyuan-Large by Tencent is a 389B param MOE (52B active). It's the largest open-weights MOE. In some benchmarks it exceeds Llama 3.1 405B. With MLX's new 3-bit quant it just barely fits on a single 192GB M2 Ultra! And runs at a very decent >15 toks/sec:

Awni Hannun

64,939 views • 1 year ago

Wanna know whether different LLM providers serve the same LLama 3.1 70B? I sure did! So I ran a quick eval to get some surprising results + open sourced my code 👇 Check out my comparison between Groq Inc Fireworks AI octoaicloud DeepInfra and Together AI

Wanna know whether different LLM providers serve the same LLama 3.1 70B? I sure did! So I ran a quick eval to get some surprising results + open sourced my code 👇 Check out my comparison between Groq Inc Fireworks AI octoaicloud DeepInfra and Together AI

Alex Volkov (Thursd/AI)

26,549 views • 2 years ago

Exclusive: Meta just released Llama 3.1 405B — the first-ever open-sourced frontier AI model, beating top closed models like GPT-4o across several benchmarks. I sat down with Mark Zuckerberg, diving into why this marks a major moment in AI history. Timestamps: 00:00 Intro 00:38 Meta’s Llama 3.1 rundown 03:44 Real-world use cases for Llama 3.1 06:15 Educating developers on open-source AI tools 09:43 Societal implications of open-source AI 13:00 Balancing power and managing bad actors 14:40 Open source and global competition 16:59 Accelerating innovation and economic growth 20:04 Zuck on Apple and lessons from the past 24:22 Future of AI: Llama 3 and beyond 26:43 Prediction: Billions of personalized AI agents 31:32 Factors to changing anti-AI sentiment

Exclusive: Meta just released Llama 3.1 405B — the first-ever open-sourced frontier AI model, beating top closed models like GPT-4o across several benchmarks. I sat down with Mark Zuckerberg, diving into why this marks a major moment in AI history. Timestamps: 00:00 Intro 00:38 Meta’s Llama 3.1 rundown 03:44 Real-world use cases for Llama 3.1 06:15 Educating developers on open-source AI tools 09:43 Societal implications of open-source AI 13:00 Balancing power and managing bad actors 14:40 Open source and global competition 16:59 Accelerating innovation and economic growth 20:04 Zuck on Apple and lessons from the past 24:22 Future of AI: Llama 3 and beyond 26:43 Prediction: Billions of personalized AI agents 31:32 Factors to changing anti-AI sentiment

Rowan Cheung

2,652,458 views • 2 years ago

I'm about to save you $250/year in AI coding costs. Qwen 2.5 Coder just dropped yesterday. It's completely free. And it builds ranking websites in 2 minutes. Here's proof 👇 - Beats GPT-4 in coding benchmarks - Creates full websites instantly - Builds SEO tools that rank - Zero learning curve - 100% free forever Drop "BUILD" below for a free 1:1 technical SEO blueprint.

Sensitive content

I'm about to save you $250/year in AI coding costs. Qwen 2.5 Coder just dropped yesterday. It's completely free. And it builds ranking websites in 2 minutes. Here's proof 👇 - Beats GPT-4 in coding benchmarks - Creates full websites instantly - Builds SEO tools that rank - Zero learning curve - 100% free forever Drop "BUILD" below for a free 1:1 technical SEO blueprint.

Julian Goldie SEO

12,989 views • 1 year ago

BOOM! My early testing of FREE Open Source LOCALLY RUN Qwen-2.5 AI Models shows it is already beating DeepSeek across all categories and OpenAI. It is full multimodal and can accept Video, audio, text in and out. Week is not over yet… Here is some of the demonstration video:

BOOM! My early testing of FREE Open Source LOCALLY RUN Qwen-2.5 AI Models shows it is already beating DeepSeek across all categories and OpenAI. It is full multimodal and can accept Video, audio, text in and out. Week is not over yet… Here is some of the demonstration video:

Brian Roemmele

132,331 views • 1 year ago

🚨Breaking: Kimi K2 model is out and available on the web. Official announcement may drop anytime, and according to my sources, it will be open source too. It’s designed for coding and agentic tasks.

🚨Breaking: Kimi K2 model is out and available on the web. Official announcement may drop anytime, and according to my sources, it will be open source too. It’s designed for coding and agentic tasks.

AshutoshShrivastava

55,824 views • 1 year ago

You can run Hermes agent by Nous Research with Kimi K2.6 on atomicbot.ai VPS 🌑 Kimi.ai just dropped a new open source coding model. We asked Hermes to build and deploy a game. It did incredibly well!

You can run Hermes agent by Nous Research with Kimi K2.6 on atomicbot.ai VPS 🌑 Kimi.ai just dropped a new open source coding model. We asked Hermes to build and deploy a game. It did incredibly well!

atomicbot.ai

42,913 views • 3 months ago

Friendly reminder that you can run Qwen 3.5 (and other open-source models) on your phone And the app (PocketPal) is: - Completely free - 100% open-source - Available on iOS and Android You can then chat even if you're offline or just don't want to share personal data. Links below

Friendly reminder that you can run Qwen 3.5 (and other open-source models) on your phone And the app (PocketPal) is: - Completely free - 100% open-source - Available on iOS and Android You can then chat even if you're offline or just don't want to share personal data. Links below

Paul Couvert

26,037 views • 4 months ago