Loktar 🇺🇸's banner

Loktar 🇺🇸

@loktar00 • 7,121 subscribers

I test what models can build. Christian. Dad. Veteran.

Shorts

Qwen 3.6 27b Bf16 vs Q4 Not cherry picked first result from each, the prompt is from 𝗭𝗲𝗻 𝗠𝗮𝗴𝗻𝗲𝘁𝘀 check out his pinned post it's a great test! I did throw Q4 a bone and gen twice more.. 2nd was also all white and the third was an error. Qwen 27b has a lot of life in it still!

Qwen 3.6 27b Bf16 vs Q4 Not cherry picked first result from each, the prompt is from 𝗭𝗲𝗻 𝗠𝗮𝗴𝗻𝗲𝘁𝘀 check out his pinned post it's a great test! I did throw Q4 a bone and gen twice more.. 2nd was also all white and the third was an error. Qwen 27b has a lot of life in it still!

136,659 次观看

Now at 256k context and 62 tk/s for IQ4 DSpark, no significant degradation I can see between IQ4 and Q8. It is really cool being able to run these locally. Q8 works but is at 25.3 tk/s too slow for daily use for me anyway.

Now at 256k context and 62 tk/s for IQ4 DSpark, no significant degradation I can see between IQ4 and Q8. It is really cool being able to run these locally. Q8 works but is at 25.3 tk/s too slow for daily use for me anyway.

13,020 次观看

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

The craziest thing happened to me in an interview yesterday. I interviewed my first deep fake candidate. It took me a bit to catch on. My coworker noticed his glasses looked off.... there was a second ghost frame smeared across his cheeks. He'd been running a live face filter the whole interview So I asked him to put his hand in front of his face. He put it up right NEXT to his face lol, never in front. Asked again, still wouldn't do it He looked a little different from his linked in and his github (which I wont link here) hiring and interviewing is so ridiculous now. In the days of AI we're going to have to be more careful.. and forced to move to in person interviews.

The craziest thing happened to me in an interview yesterday. I interviewed my first deep fake candidate. It took me a bit to catch on. My coworker noticed his glasses looked off.... there was a second ghost frame smeared across his cheeks. He'd been running a live face filter the whole interview So I asked him to put his hand in front of his face. He put it up right NEXT to his face lol, never in front. Asked again, still wouldn't do it He looked a little different from his linked in and his github (which I wont link here) hiring and interviewing is so ridiculous now. In the days of AI we're going to have to be more careful.. and forced to move to in person interviews.

Loktar 🇺🇸

556,364 次观看 • 11 天前

Deepseek 4.1 Flash is such a GOAT. This isn't Opus, this is a Flash model! 1 turn it did this, it's just so freaking good! Tasked it to create a realistic grassy hill with realistic grass, trees and dandelions blowing in the wind. Even has great sound generated by the model!

Deepseek 4.1 Flash is such a GOAT. This isn't Opus, this is a Flash model! 1 turn it did this, it's just so freaking good! Tasked it to create a realistic grassy hill with realistic grass, trees and dandelions blowing in the wind. Even has great sound generated by the model!

Loktar 🇺🇸

52,311 次观看 • 11 天前

DeepSeek V4.1 Flash 1017 everyone's talking about it for a reason. It does great with 3d built this honeybee on a sunflower like it was nothing. Used Deepseek Harness, 1 turn, looks so good!

DeepSeek V4.1 Flash 1017 everyone's talking about it for a reason. It does great with 3d built this honeybee on a sunflower like it was nothing. Used Deepseek Harness, 1 turn, looks so good!

Loktar 🇺🇸

43,671 次观看 • 10 天前

This is Deepseek V4.1 Flash, are they doing a stealth upgrade behind the scenes? 1 prompt, 0 revisions, no harness straight to the API. This is the kind of output I expect and hope for from Anth/OpenAi models, NOT from an open weight Flash model. This is insane!

This is Deepseek V4.1 Flash, are they doing a stealth upgrade behind the scenes? 1 prompt, 0 revisions, no harness straight to the API. This is the kind of output I expect and hope for from Anth/OpenAi models, NOT from an open weight Flash model. This is insane!

Loktar 🇺🇸

47,768 次观看 • 15 天前

Deepseek 1017 this is the same prompt I used with Opus and a few other models.. Deepseek is definitely the best Open model result by far, rivaling Opus result.

Deepseek 1017 this is the same prompt I used with Opus and a few other models.. Deepseek is definitely the best Open model result by far, rivaling Opus result.

Loktar 🇺🇸

13,312 次观看 • 8 天前

Muse Spark 1.3 xhigh Space Station Flythrough

Muse Spark 1.3 xhigh Space Station Flythrough

Loktar 🇺🇸

36,078 次观看 • 1 个月前

Qwen 3.8 27b Makes BEUATIFUL websites!! Getting these on GH for people to look at as well. @Qwen and Qwen Developers this is such an amazing model for websites and components!

Qwen 3.8 27b Makes BEUATIFUL websites!! Getting these on GH for people to look at as well. @Qwen and Qwen Developers this is such an amazing model for websites and components!

Loktar 🇺🇸

41,362 次观看 • 2 个月前

Have you always just set your 3090s between 210 and 240 and walked away? Turns out you can drop them lower and save yourself some heat, and power! I tried all the way down to 120w

Have you always just set your 3090s between 210 and 240 and walked away? Turns out you can drop them lower and save yourself some heat, and power! I tried all the way down to 120w

Loktar 🇺🇸

17,982 次观看 • 27 天前

Deepseek 4.1 Flash is an awesome model if we can run this at home were lucky yet again. This is the first fun little game I've made with AI in a while. It's not perfect but it's fun! Recommend turning sound on. Playable link below.

Deepseek 4.1 Flash is an awesome model if we can run this at home were lucky yet again. This is the first fun little game I've made with AI in a while. It's not perfect but it's fun! Recommend turning sound on. Playable link below.

Loktar 🇺🇸

27,892 次观看 • 1 个月前

Qwen 3.8 27b Pagoda test compared to 3.8 max and 3.6. BF16 results! Looks great!

Qwen 3.8 27b Pagoda test compared to 3.8 max and 3.6. BF16 results! Looks great!

Loktar 🇺🇸

36,732 次观看 • 2 个月前

We are so spoiled with models... so spoiled.. GLM 5.3 Flash a 0 shot space corridor walkthrough. Prompt below. I know there are some not fans of 0 shots, but to me they give a strong indication of a models coding ability. I don't judge models solely on this, however I cannot help bug be wowed by results like this. If you gave any dev a spec doc (a vague one at that) and they nailed your expectations, no errors, no bugs you'd look at that developer more highly than one who needed a few tries.

We are so spoiled with models... so spoiled.. GLM 5.3 Flash a 0 shot space corridor walkthrough. Prompt below. I know there are some not fans of 0 shots, but to me they give a strong indication of a models coding ability. I don't judge models solely on this, however I cannot help bug be wowed by results like this. If you gave any dev a spec doc (a vague one at that) and they nailed your expectations, no errors, no bugs you'd look at that developer more highly than one who needed a few tries.

Loktar 🇺🇸

25,882 次观看 • 2 个月前

Qwen 3.8 Flash Next locally is actually blowing my mind!! Starting to get it dialed in a bit better, it can handle very complex instructions. Having the ability to run this locally is freaking amazing! 0 shot, 1 turn. Music generated with YuE2 Link below to try it

Qwen 3.8 Flash Next locally is actually blowing my mind!! Starting to get it dialed in a bit better, it can handle very complex instructions. Having the ability to run this locally is freaking amazing! 0 shot, 1 turn. Music generated with YuE2 Link below to try it

Loktar 🇺🇸

16,642 次观看 • 1 个月前

Ox Alpha, Qwen max/27b and DS4 Flash were told rebuild the Warcraft III menu live in the browser with OMP. Each model ran as an autonomous coding agent it analyzed my capture of the real menu (video and images), then built it over 5 turns. Which one got the closest?

Ox Alpha, Qwen max/27b and DS4 Flash were told rebuild the Warcraft III menu live in the browser with OMP. Each model ran as an autonomous coding agent it analyzed my capture of the real menu (video and images), then built it over 5 turns. Which one got the closest?

Loktar 🇺🇸

24,196 次观看 • 2 个月前

Used Qwen 3.8 27b for it all (local) with OMP. I love making silly things like this, and AI allows me to move so much quicker (and with better art). Link below.

Used Qwen 3.8 27b for it all (local) with OMP. I love making silly things like this, and AI allows me to move so much quicker (and with better art). Link below.

Loktar 🇺🇸

21,858 次观看 • 2 个月前

I'm hiring a sr. full stack developer! Python, React, TypeScript, and of COURSE AI in your development workflow. 100% remote. Must be located in the contiguous U.S. (no exceptions). We're a Christian startup building a discipleship app to advance God's kingdom. We pray before meetings, and our faith is central to what we're building. $115k to $145k/year. Interested? DM me! Deepseek Flash for the animation

I'm hiring a sr. full stack developer! Python, React, TypeScript, and of COURSE AI in your development workflow. 100% remote. Must be located in the contiguous U.S. (no exceptions). We're a Christian startup building a discipleship app to advance God's kingdom. We pray before meetings, and our faith is central to what we're building. $115k to $145k/year. Interested? DM me! Deepseek Flash for the animation

Loktar 🇺🇸

10,680 次观看 • 26 天前

Hopefully deepseek-v4.1-flash-expires-on-0910 releases soon.. This is really a step up in general and will put A LOT of pressure on Qwen Flash Next. It's doing a really good job on 0 shots, camera in this is shaky but that's easy to fix.

Hopefully deepseek-v4.1-flash-expires-on-0910 releases soon.. This is really a step up in general and will put A LOT of pressure on Qwen Flash Next. It's doing a really good job on 0 shots, camera in this is shaky but that's easy to fix.

Loktar 🇺🇸

17,002 次观看 • 1 个月前

The new Deepseek Flash 4.1 is pretty great! Here's a project I'm working on on the side with the model. Might use this spec to test other models as well. Very fitting for the season we're stepping into. This was surprisingly quick to create! I have some more demos baking!

The new Deepseek Flash 4.1 is pretty great! Here's a project I'm working on on the side with the model. Might use this spec to test other models as well. Very fitting for the season we're stepping into. This was surprisingly quick to create! I have some more demos baking!

Loktar 🇺🇸

15,266 次观看 • 1 个月前

Qwen 3.8 27b NVFP4 The same web prompts, not bad, but BF16 is definitely king.

Qwen 3.8 27b NVFP4 The same web prompts, not bad, but BF16 is definitely king.

Loktar 🇺🇸

17,391 次观看 • 2 个月前

Laguna S-2.1 vs Qwen3.6 35b and 27b on my visual bench.. 93 prompts, thinking on, every demo has to actually run and be error free in the browser to count. The stats across the run 27B 85/93 (91%) 35B Q8 76/93 (82%) Laguna 73/93 (78%) Love having a new model to test!

Laguna S-2.1 vs Qwen3.6 35b and 27b on my visual bench.. 93 prompts, thinking on, every demo has to actually run and be error free in the browser to count. The stats across the run 27B 85/93 (91%) 35B Q8 76/93 (82%) Laguna 73/93 (78%) Love having a new model to test!

Loktar 🇺🇸

20,340 次观看 • 3 个月前

Fable 5.1 Warcraft III Menu test. I think it's the best so far, the UI items look fantastic, and it did a great job adding a moon and what it thought was a skull on the flag

Fable 5.1 Warcraft III Menu test. I think it's the best so far, the UI items look fantastic, and it did a great job adding a moon and what it thought was a skull on the flag

Loktar 🇺🇸

11,114 次观看 • 1 个月前