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.

Grok 4.5 launched today. It displays frontier-level shell discipline, chaining Bash commands and filtering their outputs more frequently than any other model tested.

Ramp Labs

13,809 subscribers

26,860 views • 15 days ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Grok 4.5 operates at Opus-level intelligence and Sonnet-level pricing! We tested it against Claude Opus 4.8, Sonnet 5, and GLM 5.2 on a SpaceX mission: liftoff, stage sep, orbit insertion, satellite deploy. Grok 4.5 animates it more accurately and vividly than Sonnet 5 and GLM 5.2. Opus is the fastest and most expensive. Grok burns more tokens but still lands ~40% cheaper than Opus, and a bit under Sonnet 5. Time / tokens / cost: Grok 4.5 — / 35k / $0.69 Opus 4.8 — / 13.7k / $1.10 Sonnet 5 — / 22.1k / $0.78 GLM 5.2 — / 92.7k / $0.13 yet GLM did the whole mission for $0.13 😮

Grok 4.5 operates at Opus-level intelligence and Sonnet-level pricing! We tested it against Claude Opus 4.8, Sonnet 5, and GLM 5.2 on a SpaceX mission: liftoff, stage sep, orbit insertion, satellite deploy. Grok 4.5 animates it more accurately and vividly than Sonnet 5 and GLM 5.2. Opus is the fastest and most expensive. Grok burns more tokens but still lands ~40% cheaper than Opus, and a bit under Sonnet 5. Time / tokens / cost: Grok 4.5 — / 35k / $0.69 Opus 4.8 — / 13.7k / $1.10 Sonnet 5 — / 22.1k / $0.78 GLM 5.2 — / 92.7k / $0.13 yet GLM did the whole mission for $0.13 😮

GMI Cloud

150,322 views • 15 days ago

🚨 Frontier Model Router - Mix And Match All the Frontier LLMs Our most popular custom router is our frontier-model router that combines the best LLMs for all your tasks GPT 5.6 sol - data analysis Fable / Opus 4.8 - coding Flash 3.5 - summarization Grok 4.5 - real-time Pick your favorite LLMs for all your tasks. As of today the custom routers can be used in ANY harness - Abacus AI, Claude, Codex, Open-Code....

🚨 Frontier Model Router - Mix And Match All the Frontier LLMs Our most popular custom router is our frontier-model router that combines the best LLMs for all your tasks GPT 5.6 sol - data analysis Fable / Opus 4.8 - coding Flash 3.5 - summarization Grok 4.5 - real-time Pick your favorite LLMs for all your tasks. As of today the custom routers can be used in ANY harness - Abacus AI, Claude, Codex, Open-Code....

Bindu Reddy

4,582,654 views • 9 days ago

Grok 4.5 might be the BEST model to run inside Hermes or OpenClaw RIGHT NOW. I've been sleeping on Grok to be honestNot anymore. It's more than 60% cheaper than Opus 4.8 and lands around $2.49 per task versus ~$12 for Fable in Claude Code. And it's fast. So what happens when you give Hermes + Grok 4.5 its own email, its own phone number, its own debit card, and access to every tool you use? You pretty much get an AI co-founder. Everything you need to know about Grok 4.5 + Hermes below. Full episode is available to watch at The Startup Ideas Podcast (SIP) 🧃 ( thanks Nick Vasilescu for coming on) I slept on Grok. Not sleeping on it anymore. Watch

Grok 4.5 might be the BEST model to run inside Hermes or OpenClaw RIGHT NOW. I've been sleeping on Grok to be honestNot anymore. It's more than 60% cheaper than Opus 4.8 and lands around $2.49 per task versus ~$12 for Fable in Claude Code. And it's fast. So what happens when you give Hermes + Grok 4.5 its own email, its own phone number, its own debit card, and access to every tool you use? You pretty much get an AI co-founder. Everything you need to know about Grok 4.5 + Hermes below. Full episode is available to watch at The Startup Ideas Podcast (SIP) 🧃 ( thanks Nick Vasilescu for coming on) I slept on Grok. Not sleeping on it anymore. Watch

GREG ISENBERG

251,882 views • 13 days ago

timelapse #160 (13.5 hrs) - grok 4.5 high in grok build was my main for today. just getting a vibe check. it understands my intent really well and never messed anything up. i do prefer to stay in the loop with it more than claude though. SpaceXAI really cooked! - not sure what it was, but i was able to stay in flow state with grok build for hours and hours. my memory and retention helped me get a lot done - exhausted 4th claude max 20x plan in 48 hours - training dspark drafters on my traces (kept gpu warm the whole time) - did grok 4.5 high sweeps and was quite impressed at how far the model has come. its around gpt 5.5 level (official post coming) - some very important meetings - contract work - optimizing amd kernels for glm 5.2 fp8 (mostly custom all-reduce whilst avoid deadlocks) for the tiny corp - one shotted and ran all frontier models on a new benchmark which ill be revealing shortly. its the toughest benchmark ive ever made. fable doesnt even come close to beating it - had grok 4.5 take over my minecraft C/CUDA implementation. felt like whack a mole but maybe i just wasnt using it right

timelapse #160 (13.5 hrs) - grok 4.5 high in grok build was my main for today. just getting a vibe check. it understands my intent really well and never messed anything up. i do prefer to stay in the loop with it more than claude though. SpaceXAI really cooked! - not sure what it was, but i was able to stay in flow state with grok build for hours and hours. my memory and retention helped me get a lot done - exhausted 4th claude max 20x plan in 48 hours - training dspark drafters on my traces (kept gpu warm the whole time) - did grok 4.5 high sweeps and was quite impressed at how far the model has come. its around gpt 5.5 level (official post coming) - some very important meetings - contract work - optimizing amd kernels for glm 5.2 fp8 (mostly custom all-reduce whilst avoid deadlocks) for the tiny corp - one shotted and ran all frontier models on a new benchmark which ill be revealing shortly. its the toughest benchmark ive ever made. fable doesnt even come close to beating it - had grok 4.5 take over my minecraft C/CUDA implementation. felt like whack a mole but maybe i just wasnt using it right

Elliot Arledge

13,619 views • 14 days ago

the thing that makes Opus 4.5 special is you can vibe code forever without it losing the plot. i vibe coded an iOS app this weekend that any frontier model could build in one shot. but the special thing about Opus is i kept adding features and fixing bugs and never oncej fell down a rabbit hole of errors or found myself with hopelessly confused spaghetti code it just kept working and working and working truly next level as a coding model

the thing that makes Opus 4.5 special is you can vibe code forever without it losing the plot. i vibe coded an iOS app this weekend that any frontier model could build in one shot. but the special thing about Opus is i kept adding features and fixing bugs and never oncej fell down a rabbit hole of errors or found myself with hopelessly confused spaghetti code it just kept working and working and working truly next level as a coding model

Dan Shipper 📧

85,253 views • 8 months ago

Thank you for all of the birthday wishes today. Sometimes it's better to give gifts than receive them, and this year more than any other, my friend needed an extra pick me up... and it was Glorious! #DIYRKO #OuttaNowhere @CiampaWWE

Thank you for all of the birthday wishes today. Sometimes it's better to give gifts than receive them, and this year more than any other, my friend needed an extra pick me up... and it was Glorious! #DIYRKO #OuttaNowhere @CiampaWWE

Johnny Gargano

302,832 views • 1 year ago

I said Grok 4.5 was a bigger deal than Fable Elon even liked my post but then GPT-5.6 Sol showed up So I gave 7 of this week's newest AI models their own computer and ran the same real tasks through Hermes Final ranking: 1. GPT-5.6 Sol: best overall 2. Grok 4.5: close second, fast and relentless 3. Muse Spark 1.1: the surprise of the week Benchmarks tell you what a model might do. Watching it control a computer shows you what it can actually do Full comparison ↓

I said Grok 4.5 was a bigger deal than Fable Elon even liked my post but then GPT-5.6 Sol showed up So I gave 7 of this week's newest AI models their own computer and ran the same real tasks through Hermes Final ranking: 1. GPT-5.6 Sol: best overall 2. Grok 4.5: close second, fast and relentless 3. Muse Spark 1.1: the surprise of the week Benchmarks tell you what a model might do. Watching it control a computer shows you what it can actually do Full comparison ↓

Nick Vasilescu

76,652 views • 14 days ago

Grovia has made over $23,000 USD. Today I'm making it free. Forever. $20,203 in template sales and $3,480 in creator commissions. It's done more than I ever expected when I first launched it. I'm happy to let more people get their hands on it. Grab it in the thread 👇

Grovia has made over $23,000 USD. Today I'm making it free. Forever. $20,203 in template sales and $3,480 in creator commissions. It's done more than I ever expected when I first launched it. I'm happy to let more people get their hands on it. Grab it in the thread 👇

Jennifer

49,787 views • 29 days ago

SkildAI, which is developing a foundational AI model for robots, has released a new video of their humanoid robot walking over objects and exploring environments it had never seen before. The robot is using vision & adapts in real time. SkildAI: "From raw images and joint feedback, the model directly outputs low-level motor commands. This single neural network enables humanoid robots to seamlessly walk across flat ground, climb stairs, and step over obstacles without any planning, mapping, or manual switching between behaviors. For all of our testing in this video, we deployed the robot in each new environment without any prior planning or mapping. We took off the humanoid’s shoes so you can hear it.

SkildAI, which is developing a foundational AI model for robots, has released a new video of their humanoid robot walking over objects and exploring environments it had never seen before. The robot is using vision & adapts in real time. SkildAI: "From raw images and joint feedback, the model directly outputs low-level motor commands. This single neural network enables humanoid robots to seamlessly walk across flat ground, climb stairs, and step over obstacles without any planning, mapping, or manual switching between behaviors. For all of our testing in this video, we deployed the robot in each new environment without any prior planning or mapping. We took off the humanoid’s shoes so you can hear it.

Sawyer Merritt

69,297 views • 11 months ago

My expectations for Tesla Robotaxi, starting today. 1. It will be cheaper than Uber or Waymo. 2. It will have a better experience than them too. How? By arriving one minute before you need it, if you order in advance. By opening doors for you automatically (only available on select models). Personalized cabin. Music, vents, seats, and much more always left where you left them. Much smoother driving. (Better than human). Coming soon, an integrated Grok experience that will make the entire experience PhD level. Meaning it can talk to you about anything at a PhD level. Quietest cabin (coming 2026 with Cybercab). And average fastest time to destination. With far more information about what is actually going on (like an accident ahead). I have a list of 1,100 Tesla fans if you want to watch their reactions to the news. What are your expectations? Will be interesting to see if any of mine will be met today. :-) Or if Grok thinks I am full of it again. Click the Grok button, let it paint in the three items it usually does, then type to it “Is Robert Scoble right?” I will do it for you in replies.

My expectations for Tesla Robotaxi, starting today. 1. It will be cheaper than Uber or Waymo. 2. It will have a better experience than them too. How? By arriving one minute before you need it, if you order in advance. By opening doors for you automatically (only available on select models). Personalized cabin. Music, vents, seats, and much more always left where you left them. Much smoother driving. (Better than human). Coming soon, an integrated Grok experience that will make the entire experience PhD level. Meaning it can talk to you about anything at a PhD level. Quietest cabin (coming 2026 with Cybercab). And average fastest time to destination. With far more information about what is actually going on (like an accident ahead). I have a list of 1,100 Tesla fans if you want to watch their reactions to the news. What are your expectations? Will be interesting to see if any of mine will be met today. :-) Or if Grok thinks I am full of it again. Click the Grok button, let it paint in the three items it usually does, then type to it “Is Robert Scoble right?” I will do it for you in replies.

Robert Scoble

210,347 views • 1 year ago

"We have to restore the state and local tax deduction (SALT). Because it punishes New York City for providing a higher level of government service than other places. And, it's completely unfair to people in New York, because New Yorkers pay more than their fair share of federal taxes."

"We have to restore the state and local tax deduction (SALT). Because it punishes New York City for providing a higher level of government service than other places. And, it's completely unfair to people in New York, because New Yorkers pay more than their fair share of federal taxes."

George Conway ⚖️🇺🇸

70,575 views • 2 months ago

🚀 AI Just Got a Major Upgrade! Say hello to Baidu ERNIE’s new reasoning model — built to push the limits of what AI can do. ✨ Sharper Factual Accuracy – delivers reliable, verified knowledge. ✨ Smarter Instruction Following – understands complex tasks with ease. ✨ Next-Level Reasoning Power – excels in coding, logic, and creative writing. Powered by the ERNIE-4.5 base and reinforced through advanced multi-round reinforcement learning, this model doesn’t just follow commands — it thinks, plans, and executes like never before. From solving real-world problems to generating innovative ideas, Wenxin Yiyan is redefining the AI frontier. 🔥 The future of AI is here →

🚀 AI Just Got a Major Upgrade! Say hello to Baidu ERNIE’s new reasoning model — built to push the limits of what AI can do. ✨ Sharper Factual Accuracy – delivers reliable, verified knowledge. ✨ Smarter Instruction Following – understands complex tasks with ease. ✨ Next-Level Reasoning Power – excels in coding, logic, and creative writing. Powered by the ERNIE-4.5 base and reinforced through advanced multi-round reinforcement learning, this model doesn’t just follow commands — it thinks, plans, and executes like never before. From solving real-world problems to generating innovative ideas, Wenxin Yiyan is redefining the AI frontier. 🔥 The future of AI is here →

Toha Khan

32,668 views • 10 months ago

For 13 consecutive years, Chicago has had more murders than any other American city. Americans want safe streets and are tired of Democrat politicians shirking their responsibility to deal with crime. These politicians should be asking for help, not turning it away.

For 13 consecutive years, Chicago has had more murders than any other American city. Americans want safe streets and are tired of Democrat politicians shirking their responsibility to deal with crime. These politicians should be asking for help, not turning it away.

Sen. James Lankford

14,872 views • 10 months ago

Good Morning Nusantara, negeri jiran, land of smiles, pearl of the orient, kingdom of wonder, land of a million elephants, land of the ascending dragon, city of a thousand pagodas — Good Morning Southeast Asia! 🌤️ Grok Imagine Grok is currently the best for physics. The dynamic motion is simply better than the rest. No question. Remember, I used to clown on this model constantly. I laughed at the over-hyped posts from their Grokfluencers (Honestly, I still do 😂 but I understand, they're just doing their job). But once I ignored the noise and actually tested the model myself and saw it objectively, the results spoke for themselves 😎👍

Good Morning Nusantara, negeri jiran, land of smiles, pearl of the orient, kingdom of wonder, land of a million elephants, land of the ascending dragon, city of a thousand pagodas — Good Morning Southeast Asia! 🌤️ Grok Imagine Grok is currently the best for physics. The dynamic motion is simply better than the rest. No question. Remember, I used to clown on this model constantly. I laughed at the over-hyped posts from their Grokfluencers (Honestly, I still do 😂 but I understand, they're just doing their job). But once I ignored the noise and actually tested the model myself and saw it objectively, the results spoke for themselves 😎👍

DAN

16,103 views • 7 months ago

TPUSA has some serious explaining to do after $8.5 MILLION goes missing into a shell company, doesn't provide ANY paper work for the money, misses IRS deadlines, and pays more this year than any other year to employees when they just had a 33% reduction in their staff. This whole situation is extremely shady. Something is very wrong, and we're going to find out. The truth will ALWAYS prevail. Candace Owens

TPUSA has some serious explaining to do after $8.5 MILLION goes missing into a shell company, doesn't provide ANY paper work for the money, misses IRS deadlines, and pays more this year than any other year to employees when they just had a 33% reduction in their staff. This whole situation is extremely shady. Something is very wrong, and we're going to find out. The truth will ALWAYS prevail. Candace Owens

The SCIF

68,828 views • 8 months ago

High blood pressure kills more people worldwide than any other risk factor. Often, people don’t know they have it until it’s too late. Accurate home devices to screen for this conditions and others could be lifesaving. We discuss today for ⁦NBC News⁩. Please share!

High blood pressure kills more people worldwide than any other risk factor. Often, people don’t know they have it until it’s too late. Accurate home devices to screen for this conditions and others could be lifesaving. We discuss today for ⁦NBC News⁩. Please share!

Dr. Vin Gupta

162,557 views • 1 year ago

Microplastics are in the air you’re breathing “A new study has found the level of contamination in the air is up to 100x higher than previously thought” - Researchers tested the air inside their homes and found an average of 528microplastics per cubic meter - In their cars, it was more than 4x higher with more than 2200 plastics per cubic meter They are the result of plastic objects such as carpets and furniture degrading. “The new research suggests we may be inhaling up to 68,000 microplastics each day

Microplastics are in the air you’re breathing “A new study has found the level of contamination in the air is up to 100x higher than previously thought” - Researchers tested the air inside their homes and found an average of 528microplastics per cubic meter - In their cars, it was more than 4x higher with more than 2200 plastics per cubic meter They are the result of plastic objects such as carpets and furniture degrading. “The new research suggests we may be inhaling up to 68,000 microplastics each day

Wall Street Apes

31,048 views • 11 months ago

Compute Wars: OpenAI vs Anthopic. Why was Opus 4.5 such a breakthrough? Anthropic got lots more compute from AWS Madison and New Carlisle sites likely more than doubling their capacity. This got Anthropic got close to OpenAI's total capacity, and probably much higher effective capacity available for new model runs. Remember that it takes 6+ months between getting capacity and releasing the model, so the extra OpenAI capacity might be aligning well with the 'spud' model rather than GPT-5.4. Unless something dramatic happens, OpenAI will pull away in terms of compute available in H2 2026, but 2027 will be close. Future years are less certain but so far OpenAI has much higher planned capacity, though can't imagine Anthropic isn't pushing as hard as they can to get lots more compute. Always watch the compute, other things matter, but any new capability breakthrough probably came from throwing more compute at it.

Compute Wars: OpenAI vs Anthopic. Why was Opus 4.5 such a breakthrough? Anthropic got lots more compute from AWS Madison and New Carlisle sites likely more than doubling their capacity. This got Anthropic got close to OpenAI's total capacity, and probably much higher effective capacity available for new model runs. Remember that it takes 6+ months between getting capacity and releasing the model, so the extra OpenAI capacity might be aligning well with the 'spud' model rather than GPT-5.4. Unless something dramatic happens, OpenAI will pull away in terms of compute available in H2 2026, but 2027 will be close. Future years are less certain but so far OpenAI has much higher planned capacity, though can't imagine Anthropic isn't pushing as hard as they can to get lots more compute. Always watch the compute, other things matter, but any new capability breakthrough probably came from throwing more compute at it.

Peter Gostev

157,413 views • 3 months ago