正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

NVIDIA sent us 2 DGX Sparks. For a while we wondered what we would do with them. The memory bandwidth is 273GB/s making it 3x slower than an M3 Ultra (819GB/s) for batch_size=1 inference. But it has 4x more FLOPS (100 TFLOPS compared to 26 TFLOPS). So we thought,... show more

Alex Cheema

48,907 subscribers

281,225 次观看 • 9 个月前 •via X (Twitter)

科学技术教育

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Linear scaling achieved with multiple DeepSeek v3.1 instances. 4x macs = 4x throughput. 2x M3 Ultra Mac Studios = 1x DeepSeek @ 14 tok/sec 4x M3 Ultra Mac Studios = 2x DeepSeek @ 28 tok/sec DeepSeek V3.1 is a 671B parameter model - so at its native 8-bit quantization, it requires ~700GB of memory to run the model. EXO puts half of the layers on each device, combining their memory. EXO uses MLX distributed with TB5 interconnect, optimized for Apple Silicon. If we need higher throughput, adding two more devices lets us serve more users at once. EXO Labs handles all of this seamlessly - adding more devices to the cluster for linear scaling as we need it. The new EXO 1.0 will be open-source soonTM

Linear scaling achieved with multiple DeepSeek v3.1 instances. 4x macs = 4x throughput. 2x M3 Ultra Mac Studios = 1x DeepSeek @ 14 tok/sec 4x M3 Ultra Mac Studios = 2x DeepSeek @ 28 tok/sec DeepSeek V3.1 is a 671B parameter model - so at its native 8-bit quantization, it requires ~700GB of memory to run the model. EXO puts half of the layers on each device, combining their memory. EXO uses MLX distributed with TB5 interconnect, optimized for Apple Silicon. If we need higher throughput, adding two more devices lets us serve more users at once. EXO Labs handles all of this seamlessly - adding more devices to the cluster for linear scaling as we need it. The new EXO 1.0 will be open-source soonTM

Matt Beton

158,485 次观看 • 10 个月前

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Awni Hannun

48,673 次观看 • 5 个月前

DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is almost sure that it will get merged.

DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is almost sure that it will get merged.

antirez

83,448 次观看 • 2 个月前

Etched is deploying two new technologies in chip design: low-voltage inference and cluster-scale memory. CEO Gavin Uberti says they'll make their chips much more power-efficient and way, way faster than today's leading GPUs. He breaks it down: "We looked at a lot of early research directions, and we realized the key things that models need are way more compute and way faster memory." "If you think about inference, there are two key parts: prefill and decode. For prefill, it's a compute-bound problem. You need to have more FLOPS, more operations per second on each of your chips." "On our GPU, the bottleneck's actually thermals. You can't really run a GPU at more than around 50% of what it could theoretically do, or it'll melt." "So we're using a new technology today called low-voltage inference to try to solve this problem. You bring the voltage of the chip down dramatically, which allows us to have way, way better efficiency in terms of how much power is drawn per unit of math, and thus fit way way more flops onto the chip..." "For decode, it's all about bandwidth. Not just bandwidth on a chip, but bandwidth across your cluster. That's why we have this technology we call cluster-scale memory. It reduces the amount of time it takes to communicate from one chip to another dramatically." "As a result we can go use all of our HBM, HBM bandwidth, SRAM, SRAM bandwidth, and our scale-up domain as a single coherent pool. And that means if you're a user, you can go get much faster tokens per second, while still keeping your costs low."

Etched is deploying two new technologies in chip design: low-voltage inference and cluster-scale memory. CEO Gavin Uberti says they'll make their chips much more power-efficient and way, way faster than today's leading GPUs. He breaks it down: "We looked at a lot of early research directions, and we realized the key things that models need are way more compute and way faster memory." "If you think about inference, there are two key parts: prefill and decode. For prefill, it's a compute-bound problem. You need to have more FLOPS, more operations per second on each of your chips." "On our GPU, the bottleneck's actually thermals. You can't really run a GPU at more than around 50% of what it could theoretically do, or it'll melt." "So we're using a new technology today called low-voltage inference to try to solve this problem. You bring the voltage of the chip down dramatically, which allows us to have way, way better efficiency in terms of how much power is drawn per unit of math, and thus fit way way more flops onto the chip..." "For decode, it's all about bandwidth. Not just bandwidth on a chip, but bandwidth across your cluster. That's why we have this technology we call cluster-scale memory. It reduces the amount of time it takes to communicate from one chip to another dramatically." "As a result we can go use all of our HBM, HBM bandwidth, SRAM, SRAM bandwidth, and our scale-up domain as a single coherent pool. And that means if you're a user, you can go get much faster tokens per second, while still keeping your costs low."

TBPN

20,404 次观看 • 17 天前

BMW M3, E92. 414 hp. From a 4.0L naturally aspirated V8. Revving to 8,400 rpm with no turbo. 0–100 km/h in around 4.6 seconds. Back then, 400 hp meant revs and sound in the 2000s. Not torque, not instant speed. This was about the experience, this car wanted you to rev the engine hard, this is a backroads car for going fast. One of the last naturally aspirated M3s. Compared to a modern M340i with maybe similar power, this feels slower, but this M3 is way better feeling and its a V8 M3, the only V8 M3 and the older M3's feel way way more alive.

BMW M3, E92. 414 hp. From a 4.0L naturally aspirated V8. Revving to 8,400 rpm with no turbo. 0–100 km/h in around 4.6 seconds. Back then, 400 hp meant revs and sound in the 2000s. Not torque, not instant speed. This was about the experience, this car wanted you to rev the engine hard, this is a backroads car for going fast. One of the last naturally aspirated M3s. Compared to a modern M340i with maybe similar power, this feels slower, but this M3 is way better feeling and its a V8 M3, the only V8 M3 and the older M3's feel way way more alive.

Autowelt

58,875 次观看 • 2 个月前

We asked James Wang why Cerebras can run large models so quickly. His answer: Inference speed is mostly a memory problem. Instead of constantly pulling weights from external memory, Cerebras splits the model across multiple wafers and pipelines the layers together. “Inference is all about memory bandwidth.” “You don’t want to store weights external to the wafer, because the second it’s outside, it becomes much slower.” “We put the weights on the inside.” “The team built a new software stack that basically lets us split the models by layer and store them layer by layer across multiple wafers.” “That allows us to never have to read from external memory, and that’s what makes it so fast.” “For context, Claude is about 100 tokens per second.”

We asked James Wang why Cerebras can run large models so quickly. His answer: Inference speed is mostly a memory problem. Instead of constantly pulling weights from external memory, Cerebras splits the model across multiple wafers and pipelines the layers together. “Inference is all about memory bandwidth.” “You don’t want to store weights external to the wafer, because the second it’s outside, it becomes much slower.” “We put the weights on the inside.” “The team built a new software stack that basically lets us split the models by layer and store them layer by layer across multiple wafers.” “That allows us to never have to read from external memory, and that’s what makes it so fast.” “For context, Claude is about 100 tokens per second.”

MTS

27,935 次观看 • 1 个月前

Open-weight MiniMax M3 filled out a US customs form from a driver's license photo For this test we deployed MiniMax M3 Q4 using MLX-VLM on a Mac Studio M3 Ultra 512GB RAM. The model was tasked with reading a scanned document and an ID card photo, then completing a declaration form Output: 736 tokens · Input: 1,847 tokens · Time: ~31s The model analyzed both inputs, streamed its reasoning, and then called three tools: write_field for text fields, mark for Yes/No checkboxes, and sign for the signature and date. It extracted the required information, mapped it to the correct fields and completed the form without any manual input

Open-weight MiniMax M3 filled out a US customs form from a driver's license photo For this test we deployed MiniMax M3 Q4 using MLX-VLM on a Mac Studio M3 Ultra 512GB RAM. The model was tasked with reading a scanned document and an ID card photo, then completing a declaration form Output: 736 tokens · Input: 1,847 tokens · Time: ~31s The model analyzed both inputs, streamed its reasoning, and then called three tools: write_field for text fields, mark for Yes/No checkboxes, and sign for the signature and date. It extracted the required information, mapped it to the correct fields and completed the form without any manual input

atomic.chat

109,369 次观看 • 1 个月前

The Hermes Agent Accelerated Business Hackathon presented by NVIDIA AI × Stripe × Nous Research starts now, for builders making agents that can earn, spend, and run real operations at any scale. Our NVIDIA integrations let your team run agents safely through NemoClaw, quickly on Nemotron 3 Ultra, and intelligently with access to their extensive agent skills. The new Stripe Skills for Hermes let your agent buy what it needs, provision its own SaaS, and pay for the services it uses. We want to see what kind of business tooling you can build on top of this foundation, whether it’s a fully automated company or a framework to accelerate enterprise functions. Prizes: 1st — $10,000 cash + NVIDIA DGX Spark + $5,000 Stripe Credits 2nd — $5,000 cash + NVIDIA DGX Spark + $3,000 Stripe Credits 3rd — $2,500 cash + NVIDIA DGX Spark + $1,000 Stripe Credits To enter: 1) Tweet a 1-3 minute demo video tagging Nous Research with a short writeup 2) Drop the link in the submissions channel: 3) Fill out the submission form: Judged by Nous Research, NVIDIA, and Stripe on usefulness, viability, and presentation. Submissions due EOD Tuesday, June 30.

The Hermes Agent Accelerated Business Hackathon presented by NVIDIA AI × Stripe × Nous Research starts now, for builders making agents that can earn, spend, and run real operations at any scale. Our NVIDIA integrations let your team run agents safely through NemoClaw, quickly on Nemotron 3 Ultra, and intelligently with access to their extensive agent skills. The new Stripe Skills for Hermes let your agent buy what it needs, provision its own SaaS, and pay for the services it uses. We want to see what kind of business tooling you can build on top of this foundation, whether it’s a fully automated company or a framework to accelerate enterprise functions. Prizes: 1st — $10,000 cash + NVIDIA DGX Spark + $5,000 Stripe Credits 2nd — $5,000 cash + NVIDIA DGX Spark + $3,000 Stripe Credits 3rd — $2,500 cash + NVIDIA DGX Spark + $1,000 Stripe Credits To enter: 1) Tweet a 1-3 minute demo video tagging Nous Research with a short writeup 2) Drop the link in the submissions channel: 3) Fill out the submission form: Judged by Nous Research, NVIDIA, and Stripe on usefulness, viability, and presentation. Submissions due EOD Tuesday, June 30.

Nous Research

1,875,110 次观看 • 1 个月前

NVIDIA CEO JENSEN HUANG: THE AI APOCALYPSE NARRATIVE IS COMPLETE NONSENSE "it's computer and software. we make it better every year because we know exactly how it works" the real disruption is developers deploying local setups to out-produce entire teams: > hardware scale - from a $249 Jetson Orin Nano Super to a $2,999 DGX Spark with 128GB of unified memory > local inference - DGX Spark runs 70B-200B models on 1 PFLOP of throughput with zero cloud GPU rental costs > database privacy - redirecting Python API clients to Ollama takes one line of code to bypass subscription bills the people who configure local AI today are going to look very far ahead of the curve by 2027 read the setup guide below 👇

NVIDIA CEO JENSEN HUANG: THE AI APOCALYPSE NARRATIVE IS COMPLETE NONSENSE "it's computer and software. we make it better every year because we know exactly how it works" the real disruption is developers deploying local setups to out-produce entire teams: > hardware scale - from a $249 Jetson Orin Nano Super to a $2,999 DGX Spark with 128GB of unified memory > local inference - DGX Spark runs 70B-200B models on 1 PFLOP of throughput with zero cloud GPU rental costs > database privacy - redirecting Python API clients to Ollama takes one line of code to bypass subscription bills the people who configure local AI today are going to look very far ahead of the curve by 2027 read the setup guide below 👇

Mr. Buzzoni

15,201 次观看 • 28 天前

Unsloth AI and NVIDIA are Revolutionizing Local LLM Fine-Tuning: From RTX Desktops to DGX Spark Fine-tune popular AI models faster with Unsloth on NVIDIA RTX AI PCs such as GeForce RTX desktops and laptops to RTX PRO workstations and the new DGX Spark to build personalized assistants for coding, creative work, and complex agentic workflows. The landscape of modern AI is shifting. We are moving away from a total reliance on massive, generalized cloud models and entering the era of local, agentic AI. Whether it is tuning a chatbot to handle hyper-specific product support or building a personal assistant that manages intricate schedules, the potential for generative AI on local hardware is boundless. However, developers face a persistent bottleneck: How do you get a Small Language Model (SLM) to punch above its weight class and respond with high accuracy for specialized tasks? The answer is Fine-Tuning, and the tool of choice is Unsloth. Unsloth provides an easy and high-speed method to customize models. Optimized for efficient, low-memory training on NVIDIA GPUs, Unsloth scales effortlessly from GeForce RTX desktops and laptop all the way to the DGX Spark, the world’s smallest AI supercomputer...... Full analysis: NVIDIA NVIDIA AI NVIDIA AIDev NVIDIAnewsroom Unsloth AI Unsloth

Unsloth AI and NVIDIA are Revolutionizing Local LLM Fine-Tuning: From RTX Desktops to DGX Spark Fine-tune popular AI models faster with Unsloth on NVIDIA RTX AI PCs such as GeForce RTX desktops and laptops to RTX PRO workstations and the new DGX Spark to build personalized assistants for coding, creative work, and complex agentic workflows. The landscape of modern AI is shifting. We are moving away from a total reliance on massive, generalized cloud models and entering the era of local, agentic AI. Whether it is tuning a chatbot to handle hyper-specific product support or building a personal assistant that manages intricate schedules, the potential for generative AI on local hardware is boundless. However, developers face a persistent bottleneck: How do you get a Small Language Model (SLM) to punch above its weight class and respond with high accuracy for specialized tasks? The answer is Fine-Tuning, and the tool of choice is Unsloth. Unsloth provides an easy and high-speed method to customize models. Optimized for efficient, low-memory training on NVIDIA GPUs, Unsloth scales effortlessly from GeForce RTX desktops and laptop all the way to the DGX Spark, the world’s smallest AI supercomputer...... Full analysis: NVIDIA NVIDIA AI NVIDIA AIDev NVIDIAnewsroom Unsloth AI Unsloth

Marktechpost AI Dev News ⚡

31,551 次观看 • 7 个月前

1-bit GLM-5.2 GGUF vs. Claude 4.8 Opus vs. GPT-5.5 We gave 3 models the same prompt and compared one-shot outputs. The 1-bit GLM-5.2 GGUF ran locally on a Mac Studio M3 Ultra with 256GB RAM at ~21.6 tok/s. Which output do you like best? GGUF:

1-bit GLM-5.2 GGUF vs. Claude 4.8 Opus vs. GPT-5.5 We gave 3 models the same prompt and compared one-shot outputs. The 1-bit GLM-5.2 GGUF ran locally on a Mac Studio M3 Ultra with 256GB RAM at ~21.6 tok/s. Which output do you like best? GGUF:

Unsloth AI

1,535,700 次观看 • 25 天前

Using Starlink and a MacBook to play Blades in the Dark with DeepSeek V4 Flash as GM running on ds4 on a remote M3 Ultra and exposed through tailscale as a service in the middle of mountains. We are living in the best era of humankind! Hey... 👋🏻 not bad the speed eh?

Using Starlink and a MacBook to play Blades in the Dark with DeepSeek V4 Flash as GM running on ds4 on a remote M3 Ultra and exposed through tailscale as a service in the middle of mountains. We are living in the best era of humankind! Hey... 👋🏻 not bad the speed eh?

Ivan Fioravanti ᯅ

12,842 次观看 • 21 天前

We use FSD in an older AI3 M3, a new AI4 M3, and Cybertruck. While AI4 in the M3 may perform the best…FSD in the Cybertruck feels the coolest. 😎 Not much to see here, just another flawless FSD drive from a few weeks ago. But I thought it was a good example of how relaxing it feels inside the cabin of this beast with FSD at the helm. 😌

We use FSD in an older AI3 M3, a new AI4 M3, and Cybertruck. While AI4 in the M3 may perform the best…FSD in the Cybertruck feels the coolest. 😎 Not much to see here, just another flawless FSD drive from a few weeks ago. But I thought it was a good example of how relaxing it feels inside the cabin of this beast with FSD at the helm. 😌

Dana Chandler 🇨🇦🖖

61,667 次观看 • 11 个月前

CHINESE DEVELOPER JUST DROPPED A FULL 8-MINUTE SETUP GUIDE FOR THE NVIDIA DGX SPARK - AND SHOWED EXACTLY HOW TO RUN AI AGENTS AND ROBOTICS SIMULATIONS ON IT he unboxed the DGX Spark on camera - plugged in power, keyboard and ethernet and was inside the system within minutes then showed how to install iSECSIM 5.1 for virtual simulations, cloned the repo from GitHub, configured GCC 11 and built the entire stack from scratch 128GB of available memory, ARM processor, full Python library support and a complete AI development environment - all on a box the size of a paperback book at the end he launched a robotics environment and showed the DGX Spark handling full physics simulations in real time from unboxing to running agents in under 8 minutes - full guide on camera

CHINESE DEVELOPER JUST DROPPED A FULL 8-MINUTE SETUP GUIDE FOR THE NVIDIA DGX SPARK - AND SHOWED EXACTLY HOW TO RUN AI AGENTS AND ROBOTICS SIMULATIONS ON IT he unboxed the DGX Spark on camera - plugged in power, keyboard and ethernet and was inside the system within minutes then showed how to install iSECSIM 5.1 for virtual simulations, cloned the repo from GitHub, configured GCC 11 and built the entire stack from scratch 128GB of available memory, ARM processor, full Python library support and a complete AI development environment - all on a box the size of a paperback book at the end he launched a robotics environment and showed the DGX Spark handling full physics simulations in real time from unboxing to running agents in under 8 minutes - full guide on camera

Sprytix

57,360 次观看 • 1 个月前

"Mac Minis for example are a very good fit" - Andrej Karpathy Andrej Karpathy shouted out my work on EXO Labs in his keynote at Y Combinator AI SUS! Here's the breakdown: Right now most AI workloads run in the cloud where requests from different users are continuously batched together. These workloads are FLOPS-bound and favors hardware with the best unit economics of $ per FLOP, i.e. enterprise GPUs. The personal computing revolution will shift these workloads to personal devices with lower batch sizes (mostly batch_size=1). batch_size=1 inference is memory-bound, because all of the model parameters need to be loaded into the GPU every time a token is generated. Apple Silicon with its Unified Memory architecture has a lot of memory and memory bandwidth per $ compared to other hardware: - M4 Pro Mac Mini, 24GB @ 273GB/s, $58.33/GB, $5.13/GB/s - H100, 80GB @ 3350GB/s, $625/GB, $14.93/GB/s The unit economics of Apple Silicon are becoming more compelling with every release of the Mac. The future of AI inference looks more like open weights models (OpenAI open weights model soonTM) run at low batch_size on personal devices.

"Mac Minis for example are a very good fit" - Andrej Karpathy Andrej Karpathy shouted out my work on EXO Labs in his keynote at Y Combinator AI SUS! Here's the breakdown: Right now most AI workloads run in the cloud where requests from different users are continuously batched together. These workloads are FLOPS-bound and favors hardware with the best unit economics of $ per FLOP, i.e. enterprise GPUs. The personal computing revolution will shift these workloads to personal devices with lower batch sizes (mostly batch_size=1). batch_size=1 inference is memory-bound, because all of the model parameters need to be loaded into the GPU every time a token is generated. Apple Silicon with its Unified Memory architecture has a lot of memory and memory bandwidth per $ compared to other hardware: - M4 Pro Mac Mini, 24GB @ 273GB/s, $58.33/GB, $5.13/GB/s - H100, 80GB @ 3350GB/s, $625/GB, $14.93/GB/s The unit economics of Apple Silicon are becoming more compelling with every release of the Mac. The future of AI inference looks more like open weights models (OpenAI open weights model soonTM) run at low batch_size on personal devices.

Alex Cheema

103,566 次观看 • 1 年前

THIS AI DEVELOPER RUNS 120B PARAMETER MODELS LOCALLY AND HIS MACBOOK PRO CAN’T EVEN OPEN THEM he bought a DGX Spark, gave it the model, went back to his desk by end of day the inference was done, tested, committed his team is still waiting for the cloud GPU queue to free up the only thing between them is 128GB of unified memory in a box smaller than your lunch one Spark is a datacenter node, two Sparks is your own cluster bookmark this and send it to whoever is still paying for cloud GPUs

THIS AI DEVELOPER RUNS 120B PARAMETER MODELS LOCALLY AND HIS MACBOOK PRO CAN’T EVEN OPEN THEM he bought a DGX Spark, gave it the model, went back to his desk by end of day the inference was done, tested, committed his team is still waiting for the cloud GPU queue to free up the only thing between them is 128GB of unified memory in a box smaller than your lunch one Spark is a datacenter node, two Sparks is your own cluster bookmark this and send it to whoever is still paying for cloud GPUs

leopardracer

85,470 次观看 • 1 个月前

Mistral Vibe CLI running Devstral-small-2-2512 on M3 Ultra 512GB with LM Studio as backend 🔥 First part of the video 1x and the remaining 2x Great speed overall, 27 toks/s as average tracked on LM Studio console. Details to run it below 🧵

Mistral Vibe CLI running Devstral-small-2-2512 on M3 Ultra 512GB with LM Studio as backend 🔥 First part of the video 1x and the remaining 2x Great speed overall, 27 toks/s as average tracked on LM Studio console. Details to run it below 🧵

Ivan Fioravanti ᯅ

49,682 次观看 • 7 个月前

We are Hasmonean Racing and we have created a ‘racing family’, full of like minded people that get involved in racing for the love of the horse and the thrill of racing. We have our horses in Newmarket and have three excellent trainers. James Ferguson Racing Tom Clover Racing Brookhouse Racing We love what we do and we would love for you to join us with our next purchase. A big thanks to M3 Media Racing for putting this together for us.

We are Hasmonean Racing and we have created a ‘racing family’, full of like minded people that get involved in racing for the love of the horse and the thrill of racing. We have our horses in Newmarket and have three excellent trainers. James Ferguson Racing Tom Clover Racing Brookhouse Racing We love what we do and we would love for you to join us with our next purchase. A big thanks to M3 Media Racing for putting this together for us.

Hasmonean Racing

69,541 次观看 • 1 年前

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.

Awni Hannun

60,599 次观看 • 5 个月前

I plugged MiniMax (official) M3 into Claude Code in under 5 minutes and gave it a real task. most models need hand-holding every few steps. M3 ran 147 iterations on its own. what it actually did: → took a broken CUDA kernel → ran it for 24 hours, autonomously → pushed hardware efficiency from 7.6% to 71.3% → shipped a 9.4x speedup I didn't touch it once across that whole window. the rest of the spec sheet: → 1M token context window → native multimodal from day one → fully open weights pricing: → $20/month for 1.7B tokens → $50/month for 5.1B tokens → text, image, speech, and music all share one pool usually with this kind of capability the bill catches up fast. M3 keeps both lines low at the same time. that's the part that surprised me. and if this is what M3 does at launch, the full rollout is going to be one to watch. what would you build with 1M context? #MiniMax #MiniMaxM3 #ClaudeCode #LLM

I plugged MiniMax (official) M3 into Claude Code in under 5 minutes and gave it a real task. most models need hand-holding every few steps. M3 ran 147 iterations on its own. what it actually did: → took a broken CUDA kernel → ran it for 24 hours, autonomously → pushed hardware efficiency from 7.6% to 71.3% → shipped a 9.4x speedup I didn't touch it once across that whole window. the rest of the spec sheet: → 1M token context window → native multimodal from day one → fully open weights pricing: → $20/month for 1.7B tokens → $50/month for 5.1B tokens → text, image, speech, and music all share one pool usually with this kind of capability the bill catches up fast. M3 keeps both lines low at the same time. that's the part that surprised me. and if this is what M3 does at launch, the full rollout is going to be one to watch. what would you build with 1M context? #MiniMax #MiniMaxM3 #ClaudeCode #LLM

Nelly;

17,576 次观看 • 1 个月前