Loading video...

Video Failed to Load

Go Home

Gemma 4 E4B was able to search and cite 10+ websites, execute code to find the best answer! 🔥 You only need 6GB RAM to try this in Unsloth Studio. GitHub repo:

UnslothAI's profile picture

Unsloth AI

67,270 subscribers

101,740 views • 4 months ago •via X (Twitter)

EducationScience & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Qwen3.5-4B searched 20+ websites, cited its sources, and found the best answer! 🔥 Try this locally with just 4GB RAM via Unsloth Studio. The 4B model did this by executing tool calls + web search directly during its thinking trace.
0:12

Qwen3.5-4B searched 20+ websites, cited its sources, and found the best answer! 🔥 Try this locally with just 4GB RAM via Unsloth Studio. The 4B model did this by executing tool calls + web search directly during its thinking trace.

Unsloth AI

128,756 views • 5 months ago

We built the best codebase search to give any agent grounded information from GitHub repos. Introducing Sandbox Search. Point it at any repo, and we’ll spin up a secure coding agent in its own sandbox to do research for you. Use inside claude code, openclaw, cursor, and more.
0:14

We built the best codebase search to give any agent grounded information from GitHub repos. Introducing Sandbox Search. Point it at any repo, and we’ll spin up a secure coding agent in its own sandbox to do research for you. Use inside claude code, openclaw, cursor, and more.

Arlan

12,412 views • 4 months ago

GitHub Copilot Chat gets an update with the new Interactive Code Assistant View in Visual Studio, Quick Chat in Visual Studio Code, and an expanded context window to help provide better suggestions! 🥳
0:16

GitHub Copilot Chat gets an update with the new Interactive Code Assistant View in Visual Studio, Quick Chat in Visual Studio Code, and an expanded context window to help provide better suggestions! 🥳

GitHub

225,221 views • 3 years ago

returning to deepdream in the times of LLMs here I am using gemma 4 e4b, starting from a random image and optimizing its pixels to maximise the probability gemma captions them as a fantasy map with mountains, rivers, and castles actual AI art 😅
0:15

returning to deepdream in the times of LLMs here I am using gemma 4 e4b, starting from a random image and optimizing its pixels to maximise the probability gemma captions them as a fantasy map with mountains, rivers, and castles actual AI art 😅

Matt Henderson

33,690 views • 1 month ago

Does GitHub Copilot in Visual Studio Code have full project context? Yes! Just use @ workspace And with intent detection, Copilot will automatically include @ workspace for you. Available to everyone today with GitHub Copilot Free. Learn more:
0:11

Does GitHub Copilot in Visual Studio Code have full project context? Yes! Just use @ workspace And with intent detection, Copilot will automatically include @ workspace for you. Available to everyone today with GitHub Copilot Free. Learn more:

Visual Studio Code

92,919 views • 1 year ago

In tomorrow's VS Visual Studio Code insiders you can easily F1 to create GitHub Copilot CLI session in editor, to the side, and down below. Let us know how you like it. Pierce Boggan Evan Boyle
0:12

In tomorrow's VS Visual Studio Code insiders you can easily F1 to create GitHub Copilot CLI session in editor, to the side, and down below. Let us know how you like it. Pierce Boggan Evan Boyle

Peng Lyu

39,928 views • 6 months ago

We heard you. And we agree. In light of recent developments in physical media, GitHub is proud to announce that you can now obtain your public repo on CD-ROM. Keep it. Lend it to friends. Pass it on to your children. Your code is physically yours, forever. Until you lose it, let's be real. Order yours today.
0:13

We heard you. And we agree. In light of recent developments in physical media, GitHub is proud to announce that you can now obtain your public repo on CD-ROM. Keep it. Lend it to friends. Pass it on to your children. Your code is physically yours, forever. Until you lose it, let's be real. Order yours today.

GitHub

5,884,351 views • 1 month ago

Stop searching, start asking. 💡 The new GitHub Copilot search experience is live on GitHub Docs! 🎉 Ask a question, and Copilot will guide you to the solution. ✅ Get answers faster and build sooner. 🚀 Find out more.👇
0:12

Stop searching, start asking. 💡 The new GitHub Copilot search experience is live on GitHub Docs! 🎉 Ask a question, and Copilot will guide you to the solution. ✅ Get answers faster and build sooner. 🚀 Find out more.👇

GitHub

46,728 views • 1 year ago

Code LLaMA is now on Perplexity’s LLaMa Chat! Try asking it to write a function for you, or explain a code snippet: 🔗 This is the fastest way to try AI at Meta’s latest code-specialized LLM. With our model deployment expertise, we are able to provide you with this model less than 24 hours of it’s release. What’s next? We’ll integrate code LLaMA into Perplexity, all in service of providing you with the best answers to your most technical questions!
0:12

Code LLaMA is now on Perplexity’s LLaMa Chat! Try asking it to write a function for you, or explain a code snippet: 🔗 This is the fastest way to try AI at Meta’s latest code-specialized LLM. With our model deployment expertise, we are able to provide you with this model less than 24 hours of it’s release. What’s next? We’ll integrate code LLaMA into Perplexity, all in service of providing you with the best answers to your most technical questions!

Perplexity

274,587 views • 3 years ago

No video model is able to execute these movements, this kind of speed and action. All the others need to take notes. Kling 3.0 is mind-blowing.
0:10

No video model is able to execute these movements, this kind of speed and action. All the others need to take notes. Kling 3.0 is mind-blowing.

OscarAI

39,581 views • 6 months ago

Searching for files in large projects has never been easier 🚀 Check out the real-time search results for the #vscode project, which has thousands of files (and yes, we use Visual Studio Code to write Visual Studio Code). You can easily switch between fuzzy matching and continuous matching too!
0:16

Searching for files in large projects has never been easier 🚀 Check out the real-time search results for the #vscode project, which has thousands of files (and yes, we use Visual Studio Code to write Visual Studio Code). You can easily switch between fuzzy matching and continuous matching too!

Visual Studio Code

241,694 views • 1 year ago

This is the level of trust you need to be able to fall in love
0:12

This is the level of trust you need to be able to fall in love

𝑴𝒊𝒓𝒆𝒙

124,747 views • 1 month ago

Holy sh*t, this is f**king insane😳 i cancelled my higgsfield subscription for this a free repo with 7.9K stars dropped a full AI video studio that runs on your pc it runs on 6gb of vram, even old gpus wan 2.2, ltx-2, hunyuan video and flux built in no uploads, no subscriptions, no watermarks here is how you set it up: 1. git clone the repo 2. run the one-click install script 3. launch it and generate in your browser you will not find a FREE way to make AI Videos this year than this
0:15

Holy sh*t, this is f**king insane😳 i cancelled my higgsfield subscription for this a free repo with 7.9K stars dropped a full AI video studio that runs on your pc it runs on 6gb of vram, even old gpus wan 2.2, ltx-2, hunyuan video and flux built in no uploads, no subscriptions, no watermarks here is how you set it up: 1. git clone the repo 2. run the one-click install script 3. launch it and generate in your browser you will not find a FREE way to make AI Videos this year than this

painn

190,790 views • 11 days ago

/1 Gemma 4 31B just crushed Qwen 3.6 27B in a local LLM gamedev contest inside atomic.chat (prompt is below) Device: MacBook Pro M5 Max, 64GB RAM Results: Qwen 3.6 27B: 32 tokens/sec · 18m 04s · 33,946 tokens Gemma 4 31B: 27 tokens/sec · 3m 51s · 6,209 tokens So what is more important: tokens per second, or the quality of the final answer? Qwen made a very long response and showed more creativity and visual style. But Gemma gave a shorter, clearer, and more logical answer in much less time. In this one-shot Pac-Man gamedev contest, Gemma 4 31B was the clear winner. Its game logic was stronger: click reactions were smoother, and it handled interactions with elements like walls, ghosts, and particle effects better. But this was only one test. Maybe Qwen 3.6 27B can show better results with better settings. Open the comments, try our prompt, and share your result below.
0:16

/1 Gemma 4 31B just crushed Qwen 3.6 27B in a local LLM gamedev contest inside atomic.chat (prompt is below) Device: MacBook Pro M5 Max, 64GB RAM Results: Qwen 3.6 27B: 32 tokens/sec · 18m 04s · 33,946 tokens Gemma 4 31B: 27 tokens/sec · 3m 51s · 6,209 tokens So what is more important: tokens per second, or the quality of the final answer? Qwen made a very long response and showed more creativity and visual style. But Gemma gave a shorter, clearer, and more logical answer in much less time. In this one-shot Pac-Man gamedev contest, Gemma 4 31B was the clear winner. Its game logic was stronger: click reactions were smoother, and it handled interactions with elements like walls, ghosts, and particle effects better. But this was only one test. Maybe Qwen 3.6 27B can show better results with better settings. Open the comments, try our prompt, and share your result below.

Chubby♨️

72,566 views • 3 months ago

The only video I was able to find from UAE and Pogačar going for today’s record of Coll de Rates… Thanks to ridevelox
0:14

The only video I was able to find from UAE and Pogačar going for today’s record of Coll de Rates… Thanks to ridevelox

Lukáš Ronald Lukács

156,633 views • 8 months ago

First day in studio today for something different to add to the CV. When I was at school I was so dyslexic that I wasn’t allowed to do French because I couldn’t read or write in English let alone anything else. I left school at 16 - it wasn’t for me! Fast forward to working life and I never pushed for studio work at Sky because I was terrified that I wouldn’t be able to cope with the French content we show on the channel. Turns out you just need the right people around you to help find the workarounds and help get you there. Sky Sports Racing 🙏
0:11

First day in studio today for something different to add to the CV. When I was at school I was so dyslexic that I wasn’t allowed to do French because I couldn’t read or write in English let alone anything else. I left school at 16 - it wasn’t for me! Fast forward to working life and I never pushed for studio work at Sky because I was terrified that I wouldn’t be able to cope with the French content we show on the channel. Turns out you just need the right people around you to help find the workarounds and help get you there. Sky Sports Racing 🙏

Vanessa Binnie Ryle

22,133 views • 10 days ago

Go find yourself a wall and get better today Want to play at the highest level, you better be able to defend. There’s only 30 DH’s in the world, take pride in your defense You don’t need anybody else. Ultimately, it’s your career. How good do you want to be? #PicknPivot
0:10

Go find yourself a wall and get better today Want to play at the highest level, you better be able to defend. There’s only 30 DH’s in the world, take pride in your defense You don’t need anybody else. Ultimately, it’s your career. How good do you want to be? #PicknPivot

PicknPivot

69,581 views • 1 year ago

Open source AI is actually moving at an unhinged pace right now. I literally hadn't even finished typing up my last Gemma 4 12b benchmark notes before Google went ahead and dropped the official Quantization Aware Training (QAT) checkpoints on Hugging Face. If you missed the news, QAT basically bakes the compression directly into the training process. Instead of standard post training quantization degrading the model's reasoning capabilities, QAT trains the model with compression in mind. Unsloth is reporting near original performance at 4-bit with ~72% lower memory footprint. Details in the comments. Naturally, had to instantly pull the new GGUFs to see what a single RTX 4090 card (24 GB VRAM, Cuda 12.8, ubuntu 22) could do. i fired up llama.cpp engine again Look at these numbers: 1. Unsloth Gemma 4 26B-A4B IT (QAT Q4_K_XL) flags: ./build/bin/llama-cli -m gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf -cnv -ngl 99 -c 250000 -fa on -v VRAM Used: 19.5 GB context: 250,000 tokens decode throughput: 193 tps 2. Unsloth Gemma 4 31B IT (QAT Q4_K_XL) flags: Command: ./build/bin/llama-cli -m gemma-4-31B-it-qat-UD-Q4_K_XL.gguf -cnv -ngl 99 -c 60000 -fa on -v - VRAM Used: 23 GB (Tight, but zero system RAM spillover) - context: 60,000 tokens - decode throughput: 47 tps We are essentially watching hardware bottlenecks evaporate in real time. An update literally drops before you can finish benchmarking the previous one. What a time to be running local hardware. If you have a single rtx 3090, rtx 4090, these are the latest gemma models to try this week.
0:14

Open source AI is actually moving at an unhinged pace right now. I literally hadn't even finished typing up my last Gemma 4 12b benchmark notes before Google went ahead and dropped the official Quantization Aware Training (QAT) checkpoints on Hugging Face. If you missed the news, QAT basically bakes the compression directly into the training process. Instead of standard post training quantization degrading the model's reasoning capabilities, QAT trains the model with compression in mind. Unsloth is reporting near original performance at 4-bit with ~72% lower memory footprint. Details in the comments. Naturally, had to instantly pull the new GGUFs to see what a single RTX 4090 card (24 GB VRAM, Cuda 12.8, ubuntu 22) could do. i fired up llama.cpp engine again Look at these numbers: 1. Unsloth Gemma 4 26B-A4B IT (QAT Q4_K_XL) flags: ./build/bin/llama-cli -m gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf -cnv -ngl 99 -c 250000 -fa on -v VRAM Used: 19.5 GB context: 250,000 tokens decode throughput: 193 tps 2. Unsloth Gemma 4 31B IT (QAT Q4_K_XL) flags: Command: ./build/bin/llama-cli -m gemma-4-31B-it-qat-UD-Q4_K_XL.gguf -cnv -ngl 99 -c 60000 -fa on -v - VRAM Used: 23 GB (Tight, but zero system RAM spillover) - context: 60,000 tokens - decode throughput: 47 tps We are essentially watching hardware bottlenecks evaporate in real time. An update literally drops before you can finish benchmarking the previous one. What a time to be running local hardware. If you have a single rtx 3090, rtx 4090, these are the latest gemma models to try this week.

Alok

26,841 views • 2 months ago

Here's what happen when you try to input any cheat code from the first Tomb Raider into the second one This was used to troll gamers who wanted the infamous nude code
0:19

Here's what happen when you try to input any cheat code from the first Tomb Raider into the second one This was used to troll gamers who wanted the infamous nude code

𝐋𝐨𝐰 𝐏𝐨𝐥𝐲 𝐃𝐞𝐩𝐫𝐞𝐬𝐬𝐢𝐨𝐧

115,711 views • 3 years ago