Loading video...

Video Failed to Load

Go Home

One of the best uses for Muse Spark: generating HTML explainers. Dense design docs are painful to review. Point Muse Spark at one and get back an explainer you can click through — before/after states, the technical reasoning, all of it. Works on PRs too. Muse Spark strikes a...

orca_build's profile picture

Orca ADE

15,087 subscribers

15,629 views • 5 days ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Muse Spark 1.1 is here! Available today via API and on Of all the agentic use cases, I’m most excited about this one: going from a single video shot straight to task completion. 🚀 We’ve also polished and launched Computer Use and Video-Audio Perception for Muse Spark 1.1. Read more:
1:33

Muse Spark 1.1 is here! Available today via API and on Of all the agentic use cases, I’m most excited about this one: going from a single video shot straight to task completion. 🚀 We’ve also polished and launched Computer Use and Video-Audio Perception for Muse Spark 1.1. Read more:

Jiahui Yu

120,113 views • 2 months ago

Meta's Muse Spark 1.1 beat Gemini 3.6 at making 3D arcade games for 3x cheaper! We gave two models the same task: build three 3D tabletop games that play themselves Prompts: - An 8-ball pool break scattering the rack - An air hockey rally that ends in a goal - A foosball table where the rods play the ball Outputs: Muse Spark 1.1: 31.3K tokens, $0.026 Gemini 3.6: 28.8K tokens, $0.073 Muse Spark played all three tables clean with far more detail than Gemini, which barely rendered its backgrounds. Its air hockey glitched and its foosball lagged into a buggy still picture. Muse Spark 1.1 ran the full set perfectly for 3x less
0:22

Meta's Muse Spark 1.1 beat Gemini 3.6 at making 3D arcade games for 3x cheaper! We gave two models the same task: build three 3D tabletop games that play themselves Prompts: - An 8-ball pool break scattering the rack - An air hockey rally that ends in a goal - A foosball table where the rods play the ball Outputs: Muse Spark 1.1: 31.3K tokens, $0.026 Gemini 3.6: 28.8K tokens, $0.073 Muse Spark played all three tables clean with far more detail than Gemini, which barely rendered its backgrounds. Its air hockey glitched and its foosball lagged into a buggy still picture. Muse Spark 1.1 ran the full set perfectly for 3x less

atomic.chat

125,857 views • 1 month ago

It all starts with a spark of inspiration ✨ Bring your ideas to life and build applications entirely in natural language. Sign up for the technical preview of GitHub Spark:
1:14

It all starts with a spark of inspiration ✨ Bring your ideas to life and build applications entirely in natural language. Sign up for the technical preview of GitHub Spark:

GitHub

51,593 views • 1 year ago

AI at Meta's Muse Glimmer (30B dense, open-weights) launches with SGLang day-0 support. We got ~230 tok/s on a single RTX 5090, with NVFP4 + DFlash on. It also works out of the box on NVIDIA AI Developer RTX Pro 6000, DGX Spark, and MLX for Mac. Speed and reliability have always been the blockers for a local agent. SGLang and Muse Glimmer make a local agent that's blazing fast, always on, and gets things done. Go build and run a local agent on your own machine!🚀
0:32

AI at Meta's Muse Glimmer (30B dense, open-weights) launches with SGLang day-0 support. We got ~230 tok/s on a single RTX 5090, with NVFP4 + DFlash on. It also works out of the box on NVIDIA AI Developer RTX Pro 6000, DGX Spark, and MLX for Mac. Speed and reliability have always been the blockers for a local agent. SGLang and Muse Glimmer make a local agent that's blazing fast, always on, and gets things done. Go build and run a local agent on your own machine!🚀

LMSYS Org

43,106 views • 1 month ago

Muse Spark 1.2 supports a broad range of multimodal tasks, from turning visuals into working code to translating perception into physical action. It also brings robust audio-visual understanding to enable video-heavy workflows common in real-world enterprise use. Today, we’re sharing new evals and demos that illustrate the breadth of the model’s visual understanding and reasoning capabilities. Let’s start with a demo that shows how Muse Spark parses multimodal observations and calls tools to guide a robot to navigate in an unstructured environment to find a rubber duck. 🧵👇
0:43

Muse Spark 1.2 supports a broad range of multimodal tasks, from turning visuals into working code to translating perception into physical action. It also brings robust audio-visual understanding to enable video-heavy workflows common in real-world enterprise use. Today, we’re sharing new evals and demos that illustrate the breadth of the model’s visual understanding and reasoning capabilities. Let’s start with a demo that shows how Muse Spark parses multimodal observations and calls tools to guide a robot to navigate in an unstructured environment to find a rubber duck. 🧵👇

AI at Meta

54,160 views • 20 days ago

Muse Spark 1.3 Ultra Contributor vs Fable 5.1 xHigh Same prompt, both ran for roughly 2 hours The prompt had a self improvement rule: if the independent judges scored the result below 9.5/10, it had to keep improving and try again, What shocked me most was that Muse Spark just NEVER stopped, It went through 20 self improvement loops, launching 3 agents each time basically 60+ agent runs over 2 hours And the cost?, Less than $1, Yes, under $1… in Ultra mode ,This completely exceeded my expectations Huge credit to Meta, this thing is a work of art
0:24

Muse Spark 1.3 Ultra Contributor vs Fable 5.1 xHigh Same prompt, both ran for roughly 2 hours The prompt had a self improvement rule: if the independent judges scored the result below 9.5/10, it had to keep improving and try again, What shocked me most was that Muse Spark just NEVER stopped, It went through 20 self improvement loops, launching 3 agents each time basically 60+ agent runs over 2 hours And the cost?, Less than $1, Yes, under $1… in Ultra mode ,This completely exceeded my expectations Huge credit to Meta, this thing is a work of art

Iam_

149,256 views • 7 days ago

SPK is now available on Revolut across the EU, and UK. This expands access to Spark through one of the largest consumer financial platforms in Europe. Spark is infrastructure for on-chain credit markets, allocating capital across lending markets and liquidity venues through a rules-based framework designed for predictable liquidity and transparent risk. Today, Spark coordinates over $11B in capital powering lending and savings activity across DeFi. More on Spark:
0:22

SPK is now available on Revolut across the EU, and UK. This expands access to Spark through one of the largest consumer financial platforms in Europe. Spark is infrastructure for on-chain credit markets, allocating capital across lending markets and liquidity venues through a rules-based framework designed for predictable liquidity and transparent risk. Today, Spark coordinates over $11B in capital powering lending and savings activity across DeFi. More on Spark:

Spark

10,299 views • 4 months ago

unlimited meta muse spark 1.2 for free until december 31, 2026 😳 a tool you may not have heard of just did something insane: they released meta's latest model completely free. freebuff copied every feature in Lovable, bolt, replit, and base44, and released it for free with 0 paywalls. full setup guide: 1 / go to and click on Web 2 / create a free account 3 / select Muse 1.2 from the dropdown 4 / type in your prompt. it never even asks for a card. it's that simple. the direction coding tools are moving into are insane. soon, we might see Fable 5 completely free for everyone
0:21

unlimited meta muse spark 1.2 for free until december 31, 2026 😳 a tool you may not have heard of just did something insane: they released meta's latest model completely free. freebuff copied every feature in Lovable, bolt, replit, and base44, and released it for free with 0 paywalls. full setup guide: 1 / go to and click on Web 2 / create a free account 3 / select Muse 1.2 from the dropdown 4 / type in your prompt. it never even asks for a card. it's that simple. the direction coding tools are moving into are insane. soon, we might see Fable 5 completely free for everyone

Victor Cheng

50,062 views • 1 month ago

Absolution Track Thirteen - Thoughts of a Dying Atheist One of our most-played songs in January and February 2004. This was also the working title for a song from another Muse album... let us know if you know which one 👇🏻 Muse Absolution XX Anniversary out this Friday. Pre-order here:
0:57

Absolution Track Thirteen - Thoughts of a Dying Atheist One of our most-played songs in January and February 2004. This was also the working title for a song from another Muse album... let us know if you know which one 👇🏻 Muse Absolution XX Anniversary out this Friday. Pre-order here:

muse

88,443 views • 2 years ago

gemini 3.7 flash vs deepseek v4 pro 0813 vs muse spark 1.2 – on voxel city dioramas three models each built three crossy road-style 3d scenes – a construction site, a nyc intersection, a river with a drawbridge – as single self-contained html files the setup: Nous Research's hermes agent cli on OpenRouter, three.js skills preloaded, identical prompts per scene tasks: 1. construction site – tower crane on a working lift loop, paver laying fresh road, roller compacting it behind 2. nyc crossing – four-way intersection with a traffic light state machine, queuing cars, pedestrians crossing on the walk signal 3. river drawbridge – double-leaf bascule that lifts for tall boats, cars queuing at the barriers, animated water every scene: Three.js r185, box geometry only, a locked 20-color palette, four camera presets, and a day/night mode with bloom. one file, no build step, no assets models: Google DeepMind gemini 3.7 flash, DeepSeek v4 pro 0813, AI at Meta muse spark 1.2 muse and gemini finished every scene in two to three minutes. deepseek took 15 to 41 minutes per scene - build time, all three scenes #1 gemini 3.7 flash – 6m 43s #2 muse spark 1.2 – 7m 20s #3 deepseek v4 pro – 91m 25s - total tokens #1 muse spark 1.2 – 440,279 #2 gemini 3.7 flash – 713,855 #3 deepseek v4 pro – 20,957,568 - total price #1 muse spark 1.2 – $0.53 #2 gemini 3.7 flash – $0.56 #3 deepseek v4 pro – $4.57 - agent calls across the three builds #1 muse spark 1.2 – 12 #2 gemini 3.7 flash – 18 #3 deepseek v4 pro – 143 observations: • muse won two of the three scenes on looks with the smallest files in the test – 887 to 1,042 lines against gemini's 1,934 to 2,377. cheapest, fastest to a good frame, and shortest turned out to be the same column • deepseek burned 20.96m tokens – 29x gemini, 48x muse – across 143 agent calls. prompt caching is the only reason that cost $4.57: the cache discount absorbed roughly $30 of resent context • gemini was the only model whose files needed zero fixes to render – and the only one whose night mode is cosmetic. the sky never darkens and one camera button does nothing. clean code for a scene it never looked at follow thehype. for 24/7 ai news, analysis and breakdowns
0:45

gemini 3.7 flash vs deepseek v4 pro 0813 vs muse spark 1.2 – on voxel city dioramas three models each built three crossy road-style 3d scenes – a construction site, a nyc intersection, a river with a drawbridge – as single self-contained html files the setup: Nous Research's hermes agent cli on OpenRouter, three.js skills preloaded, identical prompts per scene tasks: 1. construction site – tower crane on a working lift loop, paver laying fresh road, roller compacting it behind 2. nyc crossing – four-way intersection with a traffic light state machine, queuing cars, pedestrians crossing on the walk signal 3. river drawbridge – double-leaf bascule that lifts for tall boats, cars queuing at the barriers, animated water every scene: Three.js r185, box geometry only, a locked 20-color palette, four camera presets, and a day/night mode with bloom. one file, no build step, no assets models: Google DeepMind gemini 3.7 flash, DeepSeek v4 pro 0813, AI at Meta muse spark 1.2 muse and gemini finished every scene in two to three minutes. deepseek took 15 to 41 minutes per scene - build time, all three scenes #1 gemini 3.7 flash – 6m 43s #2 muse spark 1.2 – 7m 20s #3 deepseek v4 pro – 91m 25s - total tokens #1 muse spark 1.2 – 440,279 #2 gemini 3.7 flash – 713,855 #3 deepseek v4 pro – 20,957,568 - total price #1 muse spark 1.2 – $0.53 #2 gemini 3.7 flash – $0.56 #3 deepseek v4 pro – $4.57 - agent calls across the three builds #1 muse spark 1.2 – 12 #2 gemini 3.7 flash – 18 #3 deepseek v4 pro – 143 observations: • muse won two of the three scenes on looks with the smallest files in the test – 887 to 1,042 lines against gemini's 1,934 to 2,377. cheapest, fastest to a good frame, and shortest turned out to be the same column • deepseek burned 20.96m tokens – 29x gemini, 48x muse – across 143 agent calls. prompt caching is the only reason that cost $4.57: the cache discount absorbed roughly $30 of resent context • gemini was the only model whose files needed zero fixes to render – and the only one whose night mode is cosmetic. the sky never darkens and one camera button does nothing. clean code for a scene it never looked at follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

29,033 views • 22 days ago

man design tools might be a little cooked... people talk a lot about html for nicer docs and explainers But its also really really good for riffing on product prototypes - this was like 10-20 fast rounds of back and forth to figure out what we wanna build before wiring it all into the app
0:41

man design tools might be a little cooked... people talk a lot about html for nicer docs and explainers But its also really really good for riffing on product prototypes - this was like 10-20 fast rounds of back and forth to figure out what we wanna build before wiring it all into the app

dex

20,137 views • 1 month ago

Can AI really replace SaaS? We asked Codex Spark to build a simple Salesforce clone. It generated a build plan, then wrote all the code in 29 seconds. And it works! Add contacts, live search, full pipeline dashboard – all unit tests passed. The same task on full Codex took ~5× longer. For rapid prototyping and self-contained work, nothing beats Spark. ✨
1:45

Can AI really replace SaaS? We asked Codex Spark to build a simple Salesforce clone. It generated a build plan, then wrote all the code in 29 seconds. And it works! Add contacts, live search, full pipeline dashboard – all unit tests passed. The same task on full Codex took ~5× longer. For rapid prototyping and self-contained work, nothing beats Spark. ✨

Cerebras

16,831 views • 5 months ago

Receipts in video, see it float at ~100-150 while coding the fluctuations were for task and context switching of the model. This thing rips through code! A Single NVIDIA AI DGX Spark ⚡️
0:32

Receipts in video, see it float at ~100-150 while coding the fluctuations were for task and context switching of the model. This thing rips through code! A Single NVIDIA AI DGX Spark ⚡️

ÆON FORGE ✨

10,822 views • 2 months ago

Is your relationship lacking that 'spark' it once had? There's a new app called Arya that works as an AI-powered intimacy "concierge" for couples. We spoke with the founder and one of their sex experts on TMZ Live!
7:44

Is your relationship lacking that 'spark' it once had? There's a new app called Arya that works as an AI-powered intimacy "concierge" for couples. We spoke with the founder and one of their sex experts on TMZ Live!

TMZ Live

25,690 views • 6 months ago

With Intelligent Model Routing, Replit picks the model for every task you send, so you can operate with quality, speed, and efficiency, at no extra cost. Here’s how it works, plus a breakdown on the difference between Free, Power and Max ↓
2:51

With Intelligent Model Routing, Replit picks the model for every task you send, so you can operate with quality, speed, and efficiency, at no extra cost. Here’s how it works, plus a breakdown on the difference between Free, Power and Max ↓

Replit ⠕

11,941 views • 13 days ago

Let's find the first pirates who will sail the Seas of Spark! This is for the most ferocious, ready for an adventure to conquer, who love fishing and PvPing Get 50k base:0x0fb07c88bc6d195c196279523957c004eb868248 and await the Founder Ships mint Also put a 🏴‍☠️ in comments to get one of 20 free Founder Ships
0:25

Let's find the first pirates who will sail the Seas of Spark! This is for the most ferocious, ready for an adventure to conquer, who love fishing and PvPing Get 50k base:0x0fb07c88bc6d195c196279523957c004eb868248 and await the Founder Ships mint Also put a 🏴‍☠️ in comments to get one of 20 free Founder Ships

Meta Alchemist

14,272 views • 2 months ago

Grok 4.5 crushed OpenAI & Meta in this test! Cost per run: GPT Sol $1.63 Grok 4.5 $2.47 Meta Muse Spark 1.1 $1.08 The three prompts: • Fruit Ninja – style slicing game • Angry Birds – style fort collapse • Crossy Road clone Grok nailed all three: clean physics, smooth playback, good visuals. GPT Sol did fine until Crossy Road, where it froze. Meta Muse Spark was cheapest and it showed: its Crossy Road lagged badly. The question isn't which model is cheapest, it's what a broken output costs you in reruns, wasted time, and things you can't ship. Cheap stops being cheap when it doesn't work.
0:30

Grok 4.5 crushed OpenAI & Meta in this test! Cost per run: GPT Sol $1.63 Grok 4.5 $2.47 Meta Muse Spark 1.1 $1.08 The three prompts: • Fruit Ninja – style slicing game • Angry Birds – style fort collapse • Crossy Road clone Grok nailed all three: clean physics, smooth playback, good visuals. GPT Sol did fine until Crossy Road, where it froze. Meta Muse Spark was cheapest and it showed: its Crossy Road lagged badly. The question isn't which model is cheapest, it's what a broken output costs you in reruns, wasted time, and things you can't ship. Cheap stops being cheap when it doesn't work.

AI/ML API

108,970 views • 2 months ago

meta muse spark 1.1 vs gpt 5.6 sol vs fable 5 vs grok 4.5 meta recently dropped muse spark 1.1 – a multimodal reasoning model from meta superintelligence labs built for agentic tasks. key facts: • 1m token context with active self-management – the model compacts its own history and keeps only the steps needed for later work • trained to orchestrate multi-agent systems: as main agent it plans and delegates to parallel subagents, as subagent it sticks to its job and knows when to escalate back • computer use trained to pick between scripting and clicking – writes automation when it's faster, clicks when it's simpler, batches actions per step • first public api from meta: the meta model api is now in preview • benchmarks: sweeps the agent column – mcp atlas 88.1 (opus 4.8: 82.2), jobbench 54.7 (opus: 48.4), humanity's last exam 62.1 (1st). loses coding – deepswe 1.1 53.3 vs gpt 5.5's 67.0, swe bench pro 61.5 vs opus's 69.2 our test – 3 prompts, single-file html, three.js, fully procedural, no assets: 1. norwegian house cantilevered over a fjord in a snowstorm – transmissive glass wall, fully modelled interior 2. beijing siheyuan courtyard house in dawn fog – instanced roof tiles, dougong brackets, glowing paper windows 3. new mexico adobe pueblo in an approaching dust storm – deep window reveals, windward grit accumulation we ran the test on AI/ML API platform results: - cost #1 muse spark 1.1 – $0.20 #2 grok 4.5 – $0.51 #3 gpt 5.6 sol – $1.93 #4 fable 5 – ~$5.20 - output tokens #1 muse spark 1.1 – 41,868 #2 gpt 5.6 sol – 49,139 #3 grok 4.5 – 64,954 #4 fable 5 – 81,849 - lines of code #1 muse spark 1.1 – 1,799 #2 gpt 5.6 sol – 2,377 #3 fable 5 – 3,088 #4 grok 4.5 – 4,216 observations: • muse spark is the cheapest of the four by a wide margin – 2.5x under grok, ~26x under fable per run. output quality tracks the price • only 7.4% of its output tokens are reasoning (3,104 of 41,868) – the model barely thinks before writing. economic, not pedantic: it commits to the first plan and ships it • the low loc is not compression, it's omission – all three prompts demanded instancing, muse spark delivered it in one muse spark's code quality – reviewed by fable 5: upsides: 1. all three files run 2. the adobe grit effect is legit – shader injection via onbeforecompile, windward faces detect storm direction through a normal-dot-wind term and darken procedurally 3. the fjord glass is real meshphysicalmaterial with transmission and ior, not a transparent quad 4. the siheyuan properly instances barrel tiles, dougong blocks and courtyard pavers downsides: 1. in the fjord file the strafe vector is negated – press a, you move right; press d, you move left. exactly the key mix-up we kept hitting with this model 2. all three files ship the model's self-doubt as comments: "// actually yaw orientation: need correct" sits above a direction vector that gets computed, abandoned and recomputed – dead vectors allocated every frame, 60 times a second 3. the siheyuan registers two separate keydown listeners, one containing an empty if-block 4. snow "accumulation" on the norway roof is a sine wobble on a scale value, not accumulation 5. "instanced snow" became 3,500 plain points. zero dispose calls anywhere pattern: minimal reasoning, minimal code, minimal price. it nails the flashy requirements – shaders, transmissive glass – and quietly drops the boring ones: instancing, controls, cleanup. you get a demo that mostly runs and a control scheme you can't trust follow thehype. for 24/7 ai news, analysis and breakdowns
1:01

meta muse spark 1.1 vs gpt 5.6 sol vs fable 5 vs grok 4.5 meta recently dropped muse spark 1.1 – a multimodal reasoning model from meta superintelligence labs built for agentic tasks. key facts: • 1m token context with active self-management – the model compacts its own history and keeps only the steps needed for later work • trained to orchestrate multi-agent systems: as main agent it plans and delegates to parallel subagents, as subagent it sticks to its job and knows when to escalate back • computer use trained to pick between scripting and clicking – writes automation when it's faster, clicks when it's simpler, batches actions per step • first public api from meta: the meta model api is now in preview • benchmarks: sweeps the agent column – mcp atlas 88.1 (opus 4.8: 82.2), jobbench 54.7 (opus: 48.4), humanity's last exam 62.1 (1st). loses coding – deepswe 1.1 53.3 vs gpt 5.5's 67.0, swe bench pro 61.5 vs opus's 69.2 our test – 3 prompts, single-file html, three.js, fully procedural, no assets: 1. norwegian house cantilevered over a fjord in a snowstorm – transmissive glass wall, fully modelled interior 2. beijing siheyuan courtyard house in dawn fog – instanced roof tiles, dougong brackets, glowing paper windows 3. new mexico adobe pueblo in an approaching dust storm – deep window reveals, windward grit accumulation we ran the test on AI/ML API platform results: - cost #1 muse spark 1.1 – $0.20 #2 grok 4.5 – $0.51 #3 gpt 5.6 sol – $1.93 #4 fable 5 – ~$5.20 - output tokens #1 muse spark 1.1 – 41,868 #2 gpt 5.6 sol – 49,139 #3 grok 4.5 – 64,954 #4 fable 5 – 81,849 - lines of code #1 muse spark 1.1 – 1,799 #2 gpt 5.6 sol – 2,377 #3 fable 5 – 3,088 #4 grok 4.5 – 4,216 observations: • muse spark is the cheapest of the four by a wide margin – 2.5x under grok, ~26x under fable per run. output quality tracks the price • only 7.4% of its output tokens are reasoning (3,104 of 41,868) – the model barely thinks before writing. economic, not pedantic: it commits to the first plan and ships it • the low loc is not compression, it's omission – all three prompts demanded instancing, muse spark delivered it in one muse spark's code quality – reviewed by fable 5: upsides: 1. all three files run 2. the adobe grit effect is legit – shader injection via onbeforecompile, windward faces detect storm direction through a normal-dot-wind term and darken procedurally 3. the fjord glass is real meshphysicalmaterial with transmission and ior, not a transparent quad 4. the siheyuan properly instances barrel tiles, dougong blocks and courtyard pavers downsides: 1. in the fjord file the strafe vector is negated – press a, you move right; press d, you move left. exactly the key mix-up we kept hitting with this model 2. all three files ship the model's self-doubt as comments: "// actually yaw orientation: need correct" sits above a direction vector that gets computed, abandoned and recomputed – dead vectors allocated every frame, 60 times a second 3. the siheyuan registers two separate keydown listeners, one containing an empty if-block 4. snow "accumulation" on the norway roof is a sine wobble on a scale value, not accumulation 5. "instanced snow" became 3,500 plain points. zero dispose calls anywhere pattern: minimal reasoning, minimal code, minimal price. it nails the flashy requirements – shaders, transmissive glass – and quietly drops the boring ones: instancing, controls, cleanup. you get a demo that mostly runs and a control scheme you can't trust follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

135,556 views • 2 months ago

Happy Birthday, Barack Obama! Throwing it back to those early days when folks were just discovering you… and somehow thought you were me. Still one of my favorite mix-ups of all time. Grateful for the hope you continue to spark in so many.
1:09

Happy Birthday, Barack Obama! Throwing it back to those early days when folks were just discovering you… and somehow thought you were me. Still one of my favorite mix-ups of all time. Grateful for the hope you continue to spark in so many.

Cory Booker

27,189 views • 1 month ago