Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

Today we’re announcing our “ThinkingCap” efficient model series with a 2× thinking token reduction on average in Qwen 3.6 27B, with up to 10x faster generation on individual examples.

Jaroslav Beck

18,202 subscribers

51,612 Aufrufe • vor 3 Monaten •via X (Twitter)

Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

26 Kommentare

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

The results across seven out-of-domain benchmarks: 🏆 Up to 10x faster in specific responses 📉 ~46% fewer reasoning tokens on average 🎯 −0.7pp accuracy difference (comparable accuracy) 🔄 Fewer reasoning loops & dead-ends ⚡ Lower latency, lower inference cost, better real-world throughput

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

I am using this improved model in my local cluster “Bob” for some time now. I’s fast and also shortening the answers without degrading the output.

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

Our goal was to get to same answer sooner with fewer tokens burned and without loosing accuracy.

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

🔗 Full technical write-up with methodology & numbers in our blog post: 🔗 HuggingFace LINK:

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

PS: Efficiency not just in open models is our core thesis since we started year ago. If you want to work on frontier of efficiency, join us!

Profilbild von Martin Havel

Martin Havelvor 3 Monaten

Replicated on Czech-language RAG (tender docs, 3090, same quants): −29–55 % thinking, 0 quality loss, and 3× lower variance in thinking length across seeds — more predictable latency in production. Fits 24GB. Well done 👏

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

🔥

Profilbild von Tom Krcha

Tom Krchavor 3 Monaten

🏎️ love it!

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

🙌🔥

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

;)

Profilbild von Chetin

Chetinvor 3 Monaten

Interesting.

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

your 🤖 Henry would love this model!

Profilbild von Ondřej Súkup

Ondřej Súkupvor 3 Monaten

Any plan for official mlx ?

Profilbild von Ddddddd

Dddddddvor 3 Monaten

Qwen 3.6 36b-A3 will work also?

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

Our training algorithm will work for most open models.

Profilbild von Jan Kozák 🇨🇿🇪🇺🇺🇦🇵🇱

Jan Kozák 🇨🇿🇪🇺🇺🇦🇵🇱vor 3 Monaten

i will wait for mtp, this is currently slower

Profilbild von Gedeon

Gedeonvor 3 Monaten

Good stuff

Profilbild von Latent Node

Latent Nodevor 3 Monaten

Great idea, similar to autothink - ?

Profilbild von Adel Bucetta

Adel Bucettavor 3 Monaten

what makes the 2x token reduction so remarkable is it's a direct result of rethinking compute and dataflow from scratch

Profilbild von Pavol Zajicek

Pavol Zajicekvor 3 Monaten

Interesting 👍 Does it support MTP?

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

Yep

Profilbild von Pavol Zajicek

Pavol Zajicekvor 3 Monaten

Can I make Beat Saber 2 with it? 🙃✌️

Profilbild von Jaroslav Beck

Jaroslav Beckvor 3 Monaten

Go ahead! ;)

Profilbild von Johnny Hu

Johnny Huvor 3 Monaten

Pitty it's at the beginning of summer holidays, tried to promote it to a community neewsfeed but everyone is gone for weeks :(. But thanks for making europe great again 😁 keep rocking

Profilbild von smeekky

smeekkyvor 3 Monaten

Great job, @JaroslavBeck ! Keep up the excellent work. You are truly inspiring, and the Czech Republic appreciates what you do more than it may seem at first glance. Thank you to your entire team.

Profilbild von Václav Knapp

Václav Knappvor 3 Monaten

Really cool guys!! 🇨🇿🇨🇿🔥 I have 4 questions below👇

Ähnliche Videos

Managed to get QWEN 3.6 27B running in Cursor with localhost (no Ngrok) at 120 - 140 TPS on 2x 3090s.

Managed to get QWEN 3.6 27B running in Cursor with localhost (no Ngrok) at 120 - 140 TPS on 2x 3090s.

Youssof Al Toukhi

21,221 Aufrufe • vor 5 Monaten

I put Qwen 3.6 27B on a B200 for the fun of it. I told Fable to apply my past optimisation principles. Result? 1000 TPS Qwen 3.6 27B NVFP4 (550 TPS average long decode) I used the same principles that got me 250 TPS on 2x 3090s. Now I need to buy a DGX Station 🤪.

I put Qwen 3.6 27B on a B200 for the fun of it. I told Fable to apply my past optimisation principles. Result? 1000 TPS Qwen 3.6 27B NVFP4 (550 TPS average long decode) I used the same principles that got me 250 TPS on 2x 3090s. Now I need to buy a DGX Station 🤪.

Youssof Al Toukhi

46,527 Aufrufe • vor 2 Monaten

I built a J-Space explorer with several curated sessions. It shows all three regions described in Anthropic's paper on J-Space: sensory, workspace and motor. Model: Qwen 3.6 27B 4-bit

I built a J-Space explorer with several curated sessions. It shows all three regions described in Anthropic's paper on J-Space: sensory, workspace and motor. Model: Qwen 3.6 27B 4-bit

WeZZard

38,567 Aufrufe • vor 3 Monaten

Vibe coding at 40,000 feet in the air at 75TPS+ on my MacBook with no WiFi. Qwen 3.6 27B on MTPLX V2 (coming soon!)

Vibe coding at 40,000 feet in the air at 75TPS+ on my MacBook with no WiFi. Qwen 3.6 27B on MTPLX V2 (coming soon!)

Youssof Al Toukhi

61,734 Aufrufe • vor 3 Monaten

You can now run Qwen 3.6 27B on <10GB of memory and retain 90% of the intelligence. Today, PrismML dropped Bonsai 27B - which takes their unique compression approach they have already successfully applied to Qwen3.6 8B and Flux2-Klein and applied it to a properly smart model. Here is it using an entire computer for deep research. Long horizon agentic calls already are strong with Qwen3.6 27B and they still are here - just on 10x less memory.

You can now run Qwen 3.6 27B on <10GB of memory and retain 90% of the intelligence. Today, PrismML dropped Bonsai 27B - which takes their unique compression approach they have already successfully applied to Qwen3.6 8B and Flux2-Klein and applied it to a properly smart model. Here is it using an entire computer for deep research. Long horizon agentic calls already are strong with Qwen3.6 27B and they still are here - just on 10x less memory.

Tim Carambat

63,449 Aufrufe • vor 3 Monaten

I tested Gemma 4 12b on my Macbook Pro M5 Max with 128GB of unified memory. It runs VERY slow for a 12b parameter model. 44 tokens/second. The lava lamp that it produced was also terrible. I would stick to Qwen 3.6 27b or Qwen 3.6 35b. Google is still way behind, even with their local models.

I tested Gemma 4 12b on my Macbook Pro M5 Max with 128GB of unified memory. It runs VERY slow for a 12b parameter model. 44 tokens/second. The lava lamp that it produced was also terrible. I would stick to Qwen 3.6 27b or Qwen 3.6 35b. Google is still way behind, even with their local models.

BridgeMind

70,626 Aufrufe • vor 4 Monaten

72+ TPS on Qwen 3.6 27B on a Macbook pro M5 max. MTPLX V2 out now! The fastest way to run models on MLX.

72+ TPS on Qwen 3.6 27B on a Macbook pro M5 max. MTPLX V2 out now! The fastest way to run models on MLX.

Youssof Al Toukhi

89,544 Aufrufe • vor 3 Monaten

100 TPS on 16 concurrent processes running a modified version of Cursor using QWEN 3.6 27B to build a 3D Bowmasters game.

100 TPS on 16 concurrent processes running a modified version of Cursor using QWEN 3.6 27B to build a 3D Bowmasters game.

Youssof Altoukhi

11,243 Aufrufe • vor 5 Monaten

LFM2-Audio just dropped! It's a 1.5B model that understands and generates both text and audio Inference 10x faster + quality on par with models 10x larger Available today on Hugging Face and our playground 🥳

LFM2-Audio just dropped! It's a 1.5B model that understands and generates both text and audio Inference 10x faster + quality on par with models 10x larger Available today on Hugging Face and our playground 🥳

Maxime Labonne

85,053 Aufrufe • vor 1 Jahr

Announcing Coffee: build and iterate on your UI 10x faster with AI ☕️👇

Announcing Coffee: build and iterate on your UI 10x faster with AI ☕️👇

Coframe

364,341 Aufrufe • vor 2 Jahren

Qwen-Image-2.1 in 4 steps is here ⚡ ViggleAI distilled Qwen-Image-2.1 into a 4-step turbo model, 6× faster, and holds up side by side with the full model ▶️ on Spaces

Qwen-Image-2.1 in 4 steps is here ⚡ ViggleAI distilled Qwen-Image-2.1 into a 4-step turbo model, 6× faster, and holds up side by side with the full model ▶️ on Spaces

Hugging Apps

96,750 Aufrufe • vor 1 Monat

Qwen3.8-27b produces beautiful results very presentable 😍 Way better than 3.6-27b by a long shot. My three.js and glsl tests are going on, but here is few examples, same prompt with same default settings all run on my 3090. qwen3.6-27b-Q5 vs qwen3.8-27b-Q5

Qwen3.8-27b produces beautiful results very presentable 😍 Way better than 3.6-27b by a long shot. My three.js and glsl tests are going on, but here is few examples, same prompt with same default settings all run on my 3090. qwen3.6-27b-Q5 vs qwen3.8-27b-Q5

AJ

93,007 Aufrufe • vor 2 Monaten

Today we’re introducing Navigator n2. It’s a frontier computer-use model, with just 27B parameters.

Today we’re introducing Navigator n2. It’s a frontier computer-use model, with just 27B parameters.

Devi Parikh

782,159 Aufrufe • vor 2 Monaten

1/ Offline computer-use models are here. Today, one running on our Mac made it all the way to hotel checkout. We ran Qwen3.8-27B from Qwen through Cua Driver. It also turned a flyer into a Calendar event and reconciled receipts. Every model token stayed on the Mac 🧵

1/ Offline computer-use models are here. Today, one running on our Mac made it all the way to hotel checkout. We ran Qwen3.8-27B from Qwen through Cua Driver. It also turned a flyer into a Calendar event and reconciled receipts. Every model token stayed on the Mac 🧵

Francesco

27,540 Aufrufe • vor 2 Monaten

Breaking: Browser Use is faster than a human⚡ Running Qwen 3.8 27B on 2x B200s with DFlash2. > speed ✅ > cost ✅ > accuracy ✅ (most of the time) Qwen 27B is better at using the internet than you are. The only bottleneck is website loading. Comment if you want us to deploy this to our cloud 👀

Breaking: Browser Use is faster than a human⚡ Running Qwen 3.8 27B on 2x B200s with DFlash2. > speed ✅ > cost ✅ > accuracy ✅ (most of the time) Qwen 27B is better at using the internet than you are. The only bottleneck is website loading. Comment if you want us to deploy this to our cloud 👀

Browser Use

1,107,892 Aufrufe • vor 2 Monaten

Orinth-1.0-35b vs Qwen-3.6-27b I've asked both to build a single html file of Tetris game. First here's the Orinth 35b I think it did a fine job. Every control works, the game looks fine visually. How did Qwen 3.6 27b did? See post #2

Orinth-1.0-35b vs Qwen-3.6-27b I've asked both to build a single html file of Tetris game. First here's the Orinth 35b I think it did a fine job. Every control works, the game looks fine visually. How did Qwen 3.6 27b did? See post #2

Mia

62,696 Aufrufe • vor 4 Monaten

A nice aspect of our new Gemini 3.6 Flash model is that it is much more token efficient than our 3.5 Flash model. Here's a side-by-side demonstration of that. Nice work by everyone who worked on this model and release!

A nice aspect of our new Gemini 3.6 Flash model is that it is much more token efficient than our 3.5 Flash model. Here's a side-by-side demonstration of that. Nice work by everyone who worked on this model and release!

Jeff Dean

212,506 Aufrufe • vor 3 Monaten

Today, we’re announcing a $300 million Series C in a mix of primary and secondary funding at a $3.5 billion valuation. We’re excited to bring on @Sequoia as our lead investor in the round, with participation from Spark Capital, Marathon, COATUE, CRV, & a16z.

Today, we’re announcing a $300 million Series C in a mix of primary and secondary funding at a $3.5 billion valuation. We’re excited to bring on @Sequoia as our lead investor in the round, with participation from Spark Capital, Marathon, COATUE, CRV, & a16z.

Mercury

130,579 Aufrufe • vor 1 Jahr

Qwopus 3.6-27b Coder I had a lot of requests to test it, so I did. I ran the same tests I’ve done on other models. It scored lower on tool-eval-bench than both Qwen 3.6 27b and Qwen 3.6 35b. It also produced broken single-file Tetris and Solar System HTML demos, even after 3 fresh attempts. In the Tetris html, the game stalled when you completed a line. It also randomly changed the pieces when moving them left or right. In the Solar System html, clicking on the sun or any planet broke the rendering — the objects just disappeared. Conclusion: The regular Qwen 3.6 27b and 35b models performed better. What do you think of Qwopus 3.6-27b Coder? Full results & RAW files: Full prompts in the posts below.

Qwopus 3.6-27b Coder I had a lot of requests to test it, so I did. I ran the same tests I’ve done on other models. It scored lower on tool-eval-bench than both Qwen 3.6 27b and Qwen 3.6 35b. It also produced broken single-file Tetris and Solar System HTML demos, even after 3 fresh attempts. In the Tetris html, the game stalled when you completed a line. It also randomly changed the pieces when moving them left or right. In the Solar System html, clicking on the sun or any planet broke the rendering — the objects just disappeared. Conclusion: The regular Qwen 3.6 27b and 35b models performed better. What do you think of Qwopus 3.6-27b Coder? Full results & RAW files: Full prompts in the posts below.

Mia

11,183 Aufrufe • vor 3 Monaten

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 Aufrufe • vor 2 Monaten