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.

We are releasing Ultramarine 1 Flash, a frontier-class model built for knowledge work. It is open-source and achieves similar performance to Fable 5, while being 60x cheaper. It was post-trained by a team of 5 for ~$300k all-in. Available on API now, with weights releasing in 2 weeks.

Daanish Khazi

6,577 subscribers

1,245,700 views • 20 hours ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

34 Comments

Daanish Khazi20 hours ago

It is on the Pareto frontier for knowledge work and built to support AI companies servicing banks, law firms, etc.

Daanish Khazi20 hours ago

This release is one of the largest open experiments in non-verifiable agentic RL, and we're open-sourcing our training code along with a detailed tech report.

Daanish Khazi20 hours ago

Through prolonged training, the model learnt professional standards: visual hierarchy, straightforward prose, etc.

Daanish Khazi20 hours ago

Ultramarine 1 Flash is a drop-in alternative to GPT & Claude for AI companies serving knowledge workers. Please visit our platform to make an API key, or download the Feather app to try it in our native knowledge work harness. We're admitting off of the waitlist on a rolling basis as we scale compute, and will tweet a link to the weights within 2 weeks.

eimen20 hours ago

wow that's wild. congrats!!

Konstantin Dunas20 hours ago

Congrats! Looks amazing :)

Chris Barber20 hours ago

LFG

Rtzr20 hours ago

Congrats!!

Documenting AGI20 hours ago

“Nobody has an AI background, nobody got a PhD, nobody has worked at OpenAI.” AND about to disrupt the AI industry - thrilling.

Movez20 hours ago

wow, this will be my new model for hard tasks, going to try right now

Maxim Karoubi20 hours ago

Game changer

Joumana20 hours ago

so good!

Whale Insider20 hours ago

🐳🐳🐳

tarun odedra20 hours ago

huge! want to get my hands on it

Makerth Runee20 hours ago

I will like to see how it can compete with GPT and Claude in the finance and trading niche. I hope you win over these big boys

Aleksandar Janca20 hours ago

similar to fable 5 on what though, the general benchmarks or the knowledge work tasks you actually post-trained it for

Andrew Zhao20 hours ago

Gratz!

said20 hours ago

testing it right now! congrats on the launch Daanish 🤗

Javier Rivero20 hours ago

this is insane, amazing work from the team!

Hongji Lee19 hours ago

Post-trained by a team of 5 for $300k. That's less than most companies spend on the offsite to pick an AI strategy.

Anatoli Kopadze20 hours ago

this is insane and opens so many doors, will definitely be watching

WallStreetBets20 hours ago

curious to see this hit wall street 👀

Insider Paper19 hours ago

The top US open model ranks behind 15 Chinese models on the Artificial Analysis Intelligence Index, so this is crazy.

CyrilXBT20 hours ago

Yessir it’s time for america’s deepseek baby

Rohan Bansal20 hours ago

incredible, congrats yall!

Aapakari20 hours ago

Interesting launch 🚀

Eddie Conk20 hours ago

this is a wild release, congrats!

Robert Sterling20 hours ago

Legends 💪 love seeing competition heat up

Historic Vids20 hours ago

Amazing what just five people can accomplish. Congrats to the team!

Codez20 hours ago

this is the next level of models for sure, congrats on launch !

Makerth Runee20 hours ago

Cool

Prathmesh20 hours ago

Damn built on glm 5.3 flash the goat

High Yield Harry20 hours ago

What happens when the model becomes sentient and starts making ig reels about quitting its job in IB...

Euro_maxxing20 hours ago

Love me a no-bs no special effect launch video

Related Videos

Today we're releasing Trinity-Large-Thinking. Available now on the Arcee API, with open weights on Hugging Face under Apache 2.0. We built it for developers and enterprises that want models they can inspect, post-train, host, distill, and own.

Today we're releasing Trinity-Large-Thinking. Available now on the Arcee API, with open weights on Hugging Face under Apache 2.0. We built it for developers and enterprises that want models they can inspect, post-train, host, distill, and own.

Arcee.ai

727,685 views • 6 months ago

We previously shared our research on Layer Skip, an end-to-end solution for accelerating LLMs from researchers at Meta FAIR. It achieves this by executing a subset of an LLM’s layers and utilizing subsequent layers for verification and correction. We’re now releasing inference code and fine-tuned checkpoints for this work. Model weights on Hugging Face ➡️ More details ➡️ We hope that releasing this work will open up new areas of experimentation and innovative new research in optimization and interpretability.

We previously shared our research on Layer Skip, an end-to-end solution for accelerating LLMs from researchers at Meta FAIR. It achieves this by executing a subset of an LLM’s layers and utilizing subsequent layers for verification and correction. We’re now releasing inference code and fine-tuned checkpoints for this work. Model weights on Hugging Face ➡️ More details ➡️ We hope that releasing this work will open up new areas of experimentation and innovative new research in optimization and interpretability.

AI at Meta

156,638 views • 1 year ago

Many of you asked for code & weights for π₀, we are happy to announce that we are releasing π₀ and pre-trained checkpoints in our new openpi repository! We tested the model on a few public robots, and we include code for you to fine-tune it yourself.

Many of you asked for code & weights for π₀, we are happy to announce that we are releasing π₀ and pre-trained checkpoints in our new openpi repository! We tested the model on a few public robots, and we include code for you to fine-tune it yourself.

Physical Intelligence

442,397 views • 1 year ago

📣 Claude Fable 5, the first in Anthropic's Mythos model class, is now generally available and rolling out in GitHub Copilot. It is designed for long-horizon, autonomous coding and knowledge-work tasks. Try it out in Visual Studio Code or the GitHub Copilot app. ⬇️

📣 Claude Fable 5, the first in Anthropic's Mythos model class, is now generally available and rolling out in GitHub Copilot. It is designed for long-horizon, autonomous coding and knowledge-work tasks. Try it out in Visual Studio Code or the GitHub Copilot app. ⬇️

GitHub

416,405 views • 4 months ago

🚀 1/7 We are thrilled to launch LLM360 — pushing the frontier of open-source & transparent LLMs! Starting with Amber (7B) & CrystalCoder (7B), we are releasing brand new pre-trained LLMs with all training code, data, and up to 360 model checkpoints. 🔗

LLM360

329,548 views • 2 years ago

You can now choose which model builds your app on Caffeine. The latest models are all in there: Opus 5.5 Fable 5.1 GPT-6 Astra GPT-6 Sol Kimi K3 Sonnet 5 DeepSeek V4.1 Flash Plan a whole new app with a frontier model, then switch to a faster, cheaper one for quick questions and small changes. It works on every plan, starting today.

You can now choose which model builds your app on Caffeine. The latest models are all in there: Opus 5.5 Fable 5.1 GPT-6 Astra GPT-6 Sol Kimi K3 Sonnet 5 DeepSeek V4.1 Flash Plan a whole new app with a frontier model, then switch to a faster, cheaper one for quick questions and small changes. It works on every plan, starting today.

caffeine

15,932 views • 20 days ago

Fable 5 is BACK on AI/ML API We gave Sonnet 5 and Fable 5 the exact same prompt and same real flight dataset of 70 airports and 435 routes pulled from flightradar. Then asked each one to turn it into a cinematic 3D globe as a single HTML file. Outputs: Sonnet 5: 9.8k tokens, $0.10 Fable 5: 15k tokens, $0.77 Sonnet 5 came in 87% cheaper. It drew the routes, but the planet underneath them barely exists: a dark wireframe with the arcs floating in space. Fable 5 built an actual Earth: textured oceans, ice caps, atmospheric glow. Mythos-class models are truly a masterpiece. Docs, guides & setup below:

Fable 5 is BACK on AI/ML API We gave Sonnet 5 and Fable 5 the exact same prompt and same real flight dataset of 70 airports and 435 routes pulled from flightradar. Then asked each one to turn it into a cinematic 3D globe as a single HTML file. Outputs: Sonnet 5: 9.8k tokens, $0.10 Fable 5: 15k tokens, $0.77 Sonnet 5 came in 87% cheaper. It drew the routes, but the planet underneath them barely exists: a dark wireframe with the arcs floating in space. Fable 5 built an actual Earth: textured oceans, ice caps, atmospheric glow. Mythos-class models are truly a masterpiece. Docs, guides & setup below:

AI/ML API

498,492 views • 3 months ago

Meet Apodex 1.1: Scaling Agentic Intelligence for Complex Work Open Source Harness: Open Weights: We’re excited to introduce Apodex 1.1, our new model family built to scale agentic intelligence for professional work. 🧠 Frontier-level intelligence for complex work Apodex 1.1 brings frontier-level agentic performance across complex professional work, scientific research, financial analysis, and deep search. 🤝 Asynchronous Agent Team Apodex 1.1 can break down complex tasks, coordinate multiple agents in parallel, continuously integrate their findings, and let you step in to guide or redirect the work at any time. 🔬 Open-source research workbench We’re open-sourcing FrontierAgent—a locally deployable research workbench for the Apodex 1.1 family, including asynchronous Agent Team. Available now: 🔹 Apodex 1.1 — our most capable frontier model, available through the Apodex online workbench 🔹 Apodex 1.1 mini — open-weight model for running complex work locally 🔹 FrontierAgent — open-source, locally deployable research workbench 🌐 Live workbench: 🦾 API platform: 📃 Paper:

Meet Apodex 1.1: Scaling Agentic Intelligence for Complex Work Open Source Harness: Open Weights: We’re excited to introduce Apodex 1.1, our new model family built to scale agentic intelligence for professional work. 🧠 Frontier-level intelligence for complex work Apodex 1.1 brings frontier-level agentic performance across complex professional work, scientific research, financial analysis, and deep search. 🤝 Asynchronous Agent Team Apodex 1.1 can break down complex tasks, coordinate multiple agents in parallel, continuously integrate their findings, and let you step in to guide or redirect the work at any time. 🔬 Open-source research workbench We’re open-sourcing FrontierAgent—a locally deployable research workbench for the Apodex 1.1 family, including asynchronous Agent Team. Available now: 🔹 Apodex 1.1 — our most capable frontier model, available through the Apodex online workbench 🔹 Apodex 1.1 mini — open-weight model for running complex work locally 🔹 FrontierAgent — open-source, locally deployable research workbench 🌐 Live workbench: 🦾 API platform: 📃 Paper:

Apodex

932,742 views • 1 month ago

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Base Labs

244,827 views • 1 month ago

Day 3 of 5 building with Claude Fable 5 before it goes paid. Today: an audit of everything you have already made. 1. Fable 5 reads back through all your past Claude chats 2. It flags what worked, what you dropped, and where value is going to waste 3. It ranks the top 10 fixes by effort against impact 4. Then it hands you a concrete plan for the top 3 5. Run that plan on a cheaper model and save Fable 5 for the thinking Comment FABLE for the full guide. Make sure you follow me so I can DM it to you. Day 4 tomorrow.

Day 3 of 5 building with Claude Fable 5 before it goes paid. Today: an audit of everything you have already made. 1. Fable 5 reads back through all your past Claude chats 2. It flags what worked, what you dropped, and where value is going to waste 3. It ranks the top 10 fixes by effort against impact 4. Then it hands you a concrete plan for the top 3 5. Run that plan on a cheaper model and save Fable 5 for the thinking Comment FABLE for the full guide. Make sure you follow me so I can DM it to you. Day 4 tomorrow.

Charlie Hills

13,617 views • 3 months ago

Wave 9 is here: a frontier model built for software engineering. Introducing our new family of models: SWE-1, SWE-1-lite, and SWE-1-mini. Based on internal evals, it has performance nearing that of frontier models from the foundation labs. Available now, only in Windsurf!

Wave 9 is here: a frontier model built for software engineering. Introducing our new family of models: SWE-1, SWE-1-lite, and SWE-1-mini. Based on internal evals, it has performance nearing that of frontier models from the foundation labs. Available now, only in Windsurf!

Devin Developers

743,594 views • 1 year ago

I’m very excited to share NVIDIA Kumo Tabular, a new family of foundation models for tabular data. Kumo Tabular establishes the new Pareto frontier across the entire accuracy–inference-time tradeoff. Just as importantly, we are releasing it openly: open weights, open-source software, and a permissive license for commercial use. HuggingFace: GitHub:

I’m very excited to share NVIDIA Kumo Tabular, a new family of foundation models for tabular data. Kumo Tabular establishes the new Pareto frontier across the entire accuracy–inference-time tradeoff. Just as importantly, we are releasing it openly: open weights, open-source software, and a permissive license for commercial use. HuggingFace: GitHub:

Jure Leskovec

127,600 views • 15 days ago

Krea 1 is now available to everyone. today, we're releasing the public beta of our first image model trained in collaboration with Black Forest Labs to offer superior aesthetic control and image quality. learn how to use it for free 👇

Krea 1 is now available to everyone. today, we're releasing the public beta of our first image model trained in collaboration with Black Forest Labs to offer superior aesthetic control and image quality. learn how to use it for free 👇

KREA AI

286,441 views • 1 year ago

We're releasing Paris 2.0, which, to our knowledge, is the world's first decentralized trained video generation model. We benchmarked it against a monolithic model trained on the same data and compute budget, and Paris 2.0 outperformed the monolithic by ~2x on FVD benchmark.

We're releasing Paris 2.0, which, to our knowledge, is the world's first decentralized trained video generation model. We benchmarked it against a monolithic model trained on the same data and compute budget, and Paris 2.0 outperformed the monolithic by ~2x on FVD benchmark.

Bidhan Roy

457,774 views • 4 months ago

A big milestone for AI on Windows: gpt‑oss‑20b is now available in Windows AI Foundry and Azure AI Foundry! You can now run this open-source, frontier-class model locally on a high-performance Windows PC. It’s fast, capable, and built for real-world use. Brought to life through collaboration across Windows, Azure, and OpenAI teams. We can’t wait to see what you create. 🚀

A big milestone for AI on Windows: gpt‑oss‑20b is now available in Windows AI Foundry and Azure AI Foundry! You can now run this open-source, frontier-class model locally on a high-performance Windows PC. It’s fast, capable, and built for real-world use. Brought to life through collaboration across Windows, Azure, and OpenAI teams. We can’t wait to see what you create. 🚀

Pavan Davuluri

80,680 views • 1 year ago

It's been just over a year since we started Meta Superintelligence Labs. Muse Spark was our first family of models, and now Muse Wonder is a much larger and more advanced class of models that brings us to the frontier. We're releasing two models today. Muse Wonder for fast frontier performance and Muse Wonder Full for the full version of the model.

It's been just over a year since we started Meta Superintelligence Labs. Muse Spark was our first family of models, and now Muse Wonder is a much larger and more advanced class of models that brings us to the frontier. We're releasing two models today. Muse Wonder for fast frontier performance and Muse Wonder Full for the full version of the model.

Mark Zuckerberg

552,163 views • 2 days ago

Releasing STUDIO 1939, my hand-painted, golden-age animation LoRA for minimax H3, open weight ! This 5-minute film? Every shot in it was generated in under 4 minutes total. Faster than it plays. The LoRA is available now. And the turbo H3 model available very soon on fal ( but you can already use it on normal endpoint )

Releasing STUDIO 1939, my hand-painted, golden-age animation LoRA for minimax H3, open weight ! This 5-minute film? Every shot in it was generated in under 4 minutes total. Faster than it plays. The LoRA is available now. And the turbo H3 model available very soon on fal ( but you can already use it on normal endpoint )

Lovis Odin

32,781 views • 1 month ago

GLM-5, the latest frontier open model from Z.ai, is available now on Modal. We partnered with to release an endpoint that will be free for a limited time.

GLM-5, the latest frontier open model from Z.ai, is available now on Modal. We partnered with to release an endpoint that will be free for a limited time.

Modal

58,785 views • 8 months ago

.L3Harris says Palantir helped them beat frontier AI models in less than 2 days, at 95% lower cost: "When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours." "The cost of our fine-tuned open source model was 95% lower than the frontier models we were using." "AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge." "We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates." " We own the model, we own the compute, we own the advantage."

.L3Harris says Palantir helped them beat frontier AI models in less than 2 days, at 95% lower cost: "When we fine-tuned open source models trained on our own data, we were able to outperform the frontier models in less than 48 hours." "The cost of our fine-tuned open source model was 95% lower than the frontier models we were using." "AI is a commodity. It's all about the data. We view our data as a corporate asset. It's our unique hard-earned knowledge." "We believe American defense companies should not be a vassal for frontier AI labs, handing over our data and institutional knowledge, hoping to rent back the intelligence it creates." " We own the model, we own the compute, we own the advantage."

Jawwwn

1,106,105 views • 1 month ago

Chamath is making one of the most important business arguments of 2026. Half of large US companies right now cannot generate returns that exceed their cost of capital, which has normalized back to its long run average of 8 to 11%. Another one in seven companies globally is stuck generating persistent returns between 1 and 5% and most businesses don't have room for error and in this environment walks every frontier AI lab saying the same thing, give us your data, your workflows, your processes and our model will make everything better. And companies by the millions said yes. What they didn't fully account for is what happens on the other side of that door. Every time an employee runs a query through a frontier model API, the prompt goes through external servers, workflows, customer data, pricing logic, internal processes, all of it transmitted through a third party. As Alex Karp said companies are spending on tokens while handing over the exact proprietary advantages that make their business worth owning. Microsoft blocked internal use of Anthropic's Claude Fable 5 but over its 30-day data retention policy and the largest software company in the world decided a frontier model's data handling was too risky for its own employees. A US government action revoked access to another frontier model for foreign nationals overnight. Now here's where the cost math becomes impossible to ignore. Deutsche Bank calculated a roughly 65x cost gap between frontier models like Claude Fable 5 at ~$3.25 per task and open-source alternatives at ~$0.05. For 90% of everyday enterprise tasks, performance is comparable. Open-weight models now match closed frontier systems on core agent tasks at roughly one-tenth the cost, a high-volume deployment that costs $250/day on Claude runs at $12/day on an open-source equivalent. Chamath Palihapitiya tested this directly by running a standard enterprise code migration task through an orchestration layer wrapping an open-source model came in 16.4x cheaper than using a frontier model directly.

Chamath is making one of the most important business arguments of 2026. Half of large US companies right now cannot generate returns that exceed their cost of capital, which has normalized back to its long run average of 8 to 11%. Another one in seven companies globally is stuck generating persistent returns between 1 and 5% and most businesses don't have room for error and in this environment walks every frontier AI lab saying the same thing, give us your data, your workflows, your processes and our model will make everything better. And companies by the millions said yes. What they didn't fully account for is what happens on the other side of that door. Every time an employee runs a query through a frontier model API, the prompt goes through external servers, workflows, customer data, pricing logic, internal processes, all of it transmitted through a third party. As Alex Karp said companies are spending on tokens while handing over the exact proprietary advantages that make their business worth owning. Microsoft blocked internal use of Anthropic's Claude Fable 5 but over its 30-day data retention policy and the largest software company in the world decided a frontier model's data handling was too risky for its own employees. A US government action revoked access to another frontier model for foreign nationals overnight. Now here's where the cost math becomes impossible to ignore. Deutsche Bank calculated a roughly 65x cost gap between frontier models like Claude Fable 5 at ~$3.25 per task and open-source alternatives at ~$0.05. For 90% of everyday enterprise tasks, performance is comparable. Open-weight models now match closed frontier systems on core agent tasks at roughly one-tenth the cost, a high-volume deployment that costs $250/day on Claude runs at $12/day on an open-source equivalent. Chamath Palihapitiya tested this directly by running a standard enterprise code migration task through an orchestration layer wrapping an open-source model came in 16.4x cheaper than using a frontier model directly.

Milk Road AI

282,090 views • 3 months ago