Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

New on DigitalOcean's Inference Engine: GLM-5.3-Flash from Z.ai. Agentic coding that rivals top-tier models, now natively multimodal for browser and computer use. 1M token context, MoE architecture, one endpoint swap to try it. Senior Developer Advocate Ryan O'Connor breaks it down, and Z.ai for Startups is there to help...

digitalocean's profile picture

DigitalOcean

226,534 subscribers

140,776 Aufrufe • vor 1 Monat •via X (Twitter)

Wissenschaft & TechnologieBildungNachrichten & Politik
LIVE

Anya Rossi• Live Now

Private livecam show

2 Kommentare

Profilbild von 安叫兽|Bird🕊️ 🔶 BNB
安叫兽|Bird🕊️ 🔶 BNBvor 1 Monat

@Zai_org @r_o_connor 1M 上下文挺猛，就看长任务里能不能稳住了

Profilbild von Fajar M Reza
Fajar M Rezavor 1 Monat

@Zai_org @r_o_connor Does one endpoint really simplify multimodal coding deployment for smaller teams?

Ähnliche Videos

.Kimi.ai K3 from Moonshot AI is now live on DigitalOcean Inference Engine. 1M-token context, native vision, built to run agentic tasks for hours. Supported on Inference Router and model synthesis to maximize intelligence per dollar. No setup, no model ops.
0:36

.Kimi.ai K3 from Moonshot AI is now live on DigitalOcean Inference Engine. 1M-token context, native vision, built to run agentic tasks for hours. Supported on Inference Router and model synthesis to maximize intelligence per dollar. No setup, no model ops.

DigitalOcean

668,900 Aufrufe • vor 2 Monaten

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇
0:20

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇

SGLang

99,455 Aufrufe • vor 1 Monat

GMI Cloud and Z.ai are bringing fast inference + frontier models around the globe. First stop: Singapore. Big congrats to all the builders at the GMI x Z.ai Agent Hackathon in 🇸🇬 100+ on-site, 16 teams shipped projects and battled it out for the top prizes with GLM 5.1, including $1500 for the grand winner. Proud to support all the builders with team! Keep building 💪
1:23

GMI Cloud and Z.ai are bringing fast inference + frontier models around the globe. First stop: Singapore. Big congrats to all the builders at the GMI x Z.ai Agent Hackathon in 🇸🇬 100+ on-site, 16 teams shipped projects and battled it out for the top prizes with GLM 5.1, including $1500 for the grand winner. Proud to support all the builders with team! Keep building 💪

GMI Cloud

13,733 Aufrufe • vor 6 Monaten

🎉 Day-0 support for in vLLM, available today in v0.23.0! Congrats to Z.ai on GLM-5.2, a flagship model built for long-horizon coding agents. ✨ 1M-token context, built to hold project-scale engineering work in a single run ✨ Tuned for long-horizon coding: large-scale implementation, automated research, and performance optimization ✨ One task can carry a full dev workflow, from requirements to a deployable product across platforms ✨ Client-side and mobile engineering, including an on-device debugging loop Try it out running it on vLLM today: 🔗
0:30

🎉 Day-0 support for in vLLM, available today in v0.23.0! Congrats to Z.ai on GLM-5.2, a flagship model built for long-horizon coding agents. ✨ 1M-token context, built to hold project-scale engineering work in a single run ✨ Tuned for long-horizon coding: large-scale implementation, automated research, and performance optimization ✨ One task can carry a full dev workflow, from requirements to a deployable product across platforms ✨ Client-side and mobile engineering, including an on-device debugging loop Try it out running it on vLLM today: 🔗

vLLM

36,413 Aufrufe • vor 4 Monaten

glm 5.3 flash is 7.5x cheaper, but 3.4x slower than gemini 3.7 flash Z.ai glm 5.3 flash – shipped aug 26, $0.07/$0.25 per 1m Google DeepMind gemini 3.7 flash – shipped aug 13, $0.38/$1.88 per 1m we put the two models on one job: write one html file that draws an animated 3d scene in the browser. no images, no downloads, and it has to look the same on every load. the setup: three scenes – a glass aquarium in a lit room, the solar system, a night city under a thunderstorm. identical brief word for word, reasoning effort high, 64k output cap. the numbers below are not the whole run. they cover the three scenes we kept – the best one per task from each model, the ones in the video. - total generation time for the three scenes #1 gemini 3.7 flash – 10m 36s #2 glm 5.3 flash – 36m 30s - tokens spent on those three scenes #1 glm 5.3 flash – 110k #2 gemini 3.7 flash – 111k - cost of those three scenes #1 glm 5.3 flash – $0.027 #2 gemini 3.7 flash – $0.202 observations: • glm's first 10 attempts: 7 blank pages. it kept inventing short random helpers and forgetting to define one of them. the fix was one line in the brief: use exactly one random helper, named rand(), and don't invent shorthands next to it. next 12 attempts: 11 alive, 0 crashes. • glm spends 66% of its output on reasoning, gemini 57%. that is the whole speed gap. • gemini's storm came back as a black rectangle in 4 of 6 runs. glm's best storm has a branching bolt, lit rain and wet asphalt – for $0.01. conclusion: same three scenes, same token spend – glm 5.3 flash billed $0.027 and took 36m 30s, gemini 3.7 flash billed $0.202 and took 10m 36s. glm wins gemini on price and made the best storm of the whole run follow thehype. for 24/7 ai news, analysis and breakdowns
0:30

glm 5.3 flash is 7.5x cheaper, but 3.4x slower than gemini 3.7 flash Z.ai glm 5.3 flash – shipped aug 26, $0.07/$0.25 per 1m Google DeepMind gemini 3.7 flash – shipped aug 13, $0.38/$1.88 per 1m we put the two models on one job: write one html file that draws an animated 3d scene in the browser. no images, no downloads, and it has to look the same on every load. the setup: three scenes – a glass aquarium in a lit room, the solar system, a night city under a thunderstorm. identical brief word for word, reasoning effort high, 64k output cap. the numbers below are not the whole run. they cover the three scenes we kept – the best one per task from each model, the ones in the video. - total generation time for the three scenes #1 gemini 3.7 flash – 10m 36s #2 glm 5.3 flash – 36m 30s - tokens spent on those three scenes #1 glm 5.3 flash – 110k #2 gemini 3.7 flash – 111k - cost of those three scenes #1 glm 5.3 flash – $0.027 #2 gemini 3.7 flash – $0.202 observations: • glm's first 10 attempts: 7 blank pages. it kept inventing short random helpers and forgetting to define one of them. the fix was one line in the brief: use exactly one random helper, named rand(), and don't invent shorthands next to it. next 12 attempts: 11 alive, 0 crashes. • glm spends 66% of its output on reasoning, gemini 57%. that is the whole speed gap. • gemini's storm came back as a black rectangle in 4 of 6 runs. glm's best storm has a branching bolt, lit rain and wet asphalt – for $0.01. conclusion: same three scenes, same token spend – glm 5.3 flash billed $0.027 and took 36m 30s, gemini 3.7 flash billed $0.202 and took 10m 36s. glm wins gemini on price and made the best storm of the whole run follow thehype. for 24/7 ai news, analysis and breakdowns

thehype.

16,041 Aufrufe • vor 1 Monat

ZAI is on fire. Z.ai Hot on the heels of GLM-4.6V, the team has dropped two more models on Hugging Face: - AutoGLM-Phone-9B: a 9-billion-parameter “foundation model for smartphones” that reads the screen and acts on your behalf. Small enough for on-device inference, big enough to feel like future. - GLM-ASR-Nano-2512: a pocket-sized (2B) speech-recognition powerhouse that beats Whisper v3 on English, Mandarin, and Cantonese - It even picks up your whisper. Grab them here: - AutoGLM-Phone-9B → - GLM-ASR-Nano-2512 →
1:50

ZAI is on fire. Z.ai Hot on the heels of GLM-4.6V, the team has dropped two more models on Hugging Face: - AutoGLM-Phone-9B: a 9-billion-parameter “foundation model for smartphones” that reads the screen and acts on your behalf. Small enough for on-device inference, big enough to feel like future. - GLM-ASR-Nano-2512: a pocket-sized (2B) speech-recognition powerhouse that beats Whisper v3 on English, Mandarin, and Cantonese - It even picks up your whisper. Grab them here: - AutoGLM-Phone-9B → - GLM-ASR-Nano-2512 →

Tiezhen WANG

45,514 Aufrufe • vor 10 Monaten

Serving GLM5.2 NVFP4 Agentic Workload with SGLang: How We Reached 500 TPS on 8xB300 at bs=1 In this deep dive, we explain how SGLang reaches 500+ tok/s/user at bs=1 on 8xB300, with 18 to 34% higher single-user interactivity within two weeks since day-0, and 6 to 11% better peak throughput at high concurrency, benchmarked on a real multi-turn agentic coding workload. Our new TopK-V2 kernel is 2.33x faster at 80K ISL, scaling to 10.17x at 1M ISL, keeping interactivity essentially flat out to 1M tokens. Part of the story is the architecture itself. GLM-5.2 applies IndexShare to its DSA layers and ships a stronger MTP head reusing IndexShare and KVShare. The rest comes from our serving optimizations. Special thanks to NVIDIA AI for the help in day-0 support of GLM-5.2 NVFP4, and to Z.ai for IndexShare in SGLang!
0:47

Serving GLM5.2 NVFP4 Agentic Workload with SGLang: How We Reached 500 TPS on 8xB300 at bs=1 In this deep dive, we explain how SGLang reaches 500+ tok/s/user at bs=1 on 8xB300, with 18 to 34% higher single-user interactivity within two weeks since day-0, and 6 to 11% better peak throughput at high concurrency, benchmarked on a real multi-turn agentic coding workload. Our new TopK-V2 kernel is 2.33x faster at 80K ISL, scaling to 10.17x at 1M ISL, keeping interactivity essentially flat out to 1M tokens. Part of the story is the architecture itself. GLM-5.2 applies IndexShare to its DSA layers and ships a stronger MTP head reusing IndexShare and KVShare. The rest comes from our serving optimizations. Special thanks to NVIDIA AI for the help in day-0 support of GLM-5.2 NVFP4, and to Z.ai for IndexShare in SGLang!

LMSYS Org

107,764 Aufrufe • vor 3 Monaten

📢 DigitalOcean's virtual conference for #developers, #startups, and business leaders is back! 🗓️ July 9, 2024 | 🕚 11am EDT Register now 🔗 #DOdeploy
1:00

📢 DigitalOcean's virtual conference for #developers, #startups, and business leaders is back! 🗓️ July 9, 2024 | 🕚 11am EDT Register now 🔗 #DOdeploy

DigitalOcean

31,854 Aufrufe • vor 2 Jahren

Google Antigravity is our new agentic development platform. It helps developers build faster by collaborating with AI agents that can autonomously operate across the editor, terminal, and browser. It uses Gemini 3 Pro 🧠 to reason about problems, Gemini 2.5 Computer Use 💻 for end-to-end execution, and Nano Banana 🍌 for image generation.
0:48

Google Antigravity is our new agentic development platform. It helps developers build faster by collaborating with AI agents that can autonomously operate across the editor, terminal, and browser. It uses Gemini 3 Pro 🧠 to reason about problems, Gemini 2.5 Computer Use 💻 for end-to-end execution, and Nano Banana 🍌 for image generation.

Google DeepMind

198,575 Aufrufe • vor 11 Monaten

Tencent just dropped Hy4 preview, and it's officially open source now. 770B total params, 49B active, 1M+ token context. three major releases in six months, and this one just landed in the top tier of open source models. → Built through Tencent's co-design loop: training data shaped with experts across software engineering, gaming, finance, and security, then built directly alongside products like WorkBuddy → In a blind test with 163 internal experts across 203 engineering tasks, it scored 2.99/4, ahead of GLM 5.3 at 2.92 and Kimi K3 at 2.94 → Benchmarks put it clearly past GLM 5.2 and neck and neck with GLM 5.3 → Pricing stayed low: $0.834/M input, $2.501/M output, $0.042/M cache hits I asked it to write a weekly status update email template for a team lead, then generate a filled-out example with placeholder projects, and export both as Word docs. one prompt, no follow-up, and it handed back both files clean and ready to send. free to try for the next two weeks through WorkBuddy: Tencent Hy Tencent AI WorkBuddy
0:22

Tencent just dropped Hy4 preview, and it's officially open source now. 770B total params, 49B active, 1M+ token context. three major releases in six months, and this one just landed in the top tier of open source models. → Built through Tencent's co-design loop: training data shaped with experts across software engineering, gaming, finance, and security, then built directly alongside products like WorkBuddy → In a blind test with 163 internal experts across 203 engineering tasks, it scored 2.99/4, ahead of GLM 5.3 at 2.92 and Kimi K3 at 2.94 → Benchmarks put it clearly past GLM 5.2 and neck and neck with GLM 5.3 → Pricing stayed low: $0.834/M input, $2.501/M output, $0.042/M cache hits I asked it to write a weekly status update email template for a team lead, then generate a filled-out example with placeholder projects, and export both as Word docs. one prompt, no follow-up, and it handed back both files clean and ready to send. free to try for the next two weeks through WorkBuddy: Tencent Hy Tencent AI WorkBuddy

D-Coder

109,317 Aufrufe • vor 1 Monat

🚨 GLM 5.3 just dropped, and here’s how to access the GLM family completely FREE. no subscription. no API credit cost & no rate limit. step 1️⃣: download the OpenCode app on your PC. it’s available on both Mac and Windows. 🔗: step 2️⃣: go to NVIDIA’s website, sign up with your email, verify your account, and generate a unified API key for free. NVIDIA offers free model endpoints for developers, and GLM 5.2 is one of the top models currently available there. 🔗: step 3️⃣: launch OpenCode, open the model selection menu, search for NVIDIA, and paste the API key you generated. also, there’s no rate limit, but traffic to the model through NVIDIA can get busy at times, so responses may occasionally be slower or temporarily unavailable. the only catch: NVIDIA hasn’t listed GLM 5.3 yet, so for now you’ll be using GLM 5.2. once GLM 5.3 gets added to NVIDIA’s model catalog, you should be able to switch to it the same way.
0:28

🚨 GLM 5.3 just dropped, and here’s how to access the GLM family completely FREE. no subscription. no API credit cost & no rate limit. step 1️⃣: download the OpenCode app on your PC. it’s available on both Mac and Windows. 🔗: step 2️⃣: go to NVIDIA’s website, sign up with your email, verify your account, and generate a unified API key for free. NVIDIA offers free model endpoints for developers, and GLM 5.2 is one of the top models currently available there. 🔗: step 3️⃣: launch OpenCode, open the model selection menu, search for NVIDIA, and paste the API key you generated. also, there’s no rate limit, but traffic to the model through NVIDIA can get busy at times, so responses may occasionally be slower or temporarily unavailable. the only catch: NVIDIA hasn’t listed GLM 5.3 yet, so for now you’ll be using GLM 5.2. once GLM 5.3 gets added to NVIDIA’s model catalog, you should be able to switch to it the same way.

m0h

101,243 Aufrufe • vor 2 Monaten

🎉 Congrats to MiniMax (official) on releasing MiniMax M3! Frontier coding and agentic capabilities, native image and video input, computer use, and a 1M-token context window, all in a single open model. At the heart of M3 is MSA, a new sparse attention architecture: instead of attending densely over the full KV cache, each query scores 128-token KV blocks and runs attention only over the top blocks. That is what makes 1M-token context practical to serve. M3 runs in vLLM with day-0 support, verified on NVIDIA and AMD hardware: ✨ MSA sparse attention with dedicated prefill and decode kernels ✨ 1M-token context serving with prefix caching and chunked prefill ✨ BF16 and MXFP8 checkpoints, with MoE backends for both Hopper and Blackwell ✨ Native multimodal input (image + video) ✨ Tool calling, reasoning parsing, and thinking-mode control for agent workloads Day-0 support like this is a true team effort. Grateful to the teams at MiniMax (official), NVIDIA AI, AI at AMD, and Inferact, and to the vLLM community for making it happen. 🙏 Deep dive into the implementation, kernel work, and deployment recipes: 🔗
0:21

🎉 Congrats to MiniMax (official) on releasing MiniMax M3! Frontier coding and agentic capabilities, native image and video input, computer use, and a 1M-token context window, all in a single open model. At the heart of M3 is MSA, a new sparse attention architecture: instead of attending densely over the full KV cache, each query scores 128-token KV blocks and runs attention only over the top blocks. That is what makes 1M-token context practical to serve. M3 runs in vLLM with day-0 support, verified on NVIDIA and AMD hardware: ✨ MSA sparse attention with dedicated prefill and decode kernels ✨ 1M-token context serving with prefix caching and chunked prefill ✨ BF16 and MXFP8 checkpoints, with MoE backends for both Hopper and Blackwell ✨ Native multimodal input (image + video) ✨ Tool calling, reasoning parsing, and thinking-mode control for agent workloads Day-0 support like this is a true team effort. Grateful to the teams at MiniMax (official), NVIDIA AI, AI at AMD, and Inferact, and to the vLLM community for making it happen. 🙏 Deep dive into the implementation, kernel work, and deployment recipes: 🔗

vLLM

40,306 Aufrufe • vor 4 Monaten

Muse Spark Open is built to plan, act, and use computer across agentic tasks for life and work. You can now build a powerful personal AI agent with an open stack, and SGLang to make it: - Ultra-responsive with NVFP4 MegaMoE and sequence parallelism - Faster for agentic and coding tasks with our DFlash2 draft model - Customizable for specialized use cases via fine-tuning with Miles We’re incredibly proud to have made this possible with AI at Meta, NVIDIA AI, AI at AMD, Intel, and RadixArk!
0:35

Muse Spark Open is built to plan, act, and use computer across agentic tasks for life and work. You can now build a powerful personal AI agent with an open stack, and SGLang to make it: - Ultra-responsive with NVFP4 MegaMoE and sequence parallelism - Faster for agentic and coding tasks with our DFlash2 draft model - Customizable for specialized use cases via fine-tuning with Miles We’re incredibly proud to have made this possible with AI at Meta, NVIDIA AI, AI at AMD, Intel, and RadixArk!

SGLang

285,203 Aufrufe • vor 4 Tagen

Timelapse #156 (36 hrs) - Worked with the tiny corp on getting GLM 5.2 running on 8xMI300X (sglang won here) - Launched KernelBench-Mega and updated Kernelbench-Hard with h100 and b200 sweeps - Took care of boring business stuff - Did some training sweeps for specialized technical vocab audio model - Some bugs with putting kernelbench-mega and hard on cloud instances so had to do some reruns. Learned a lot though - Setting up my own local rl infra and profiling concurrency 128 rollouts with vllm. Became clear to me that I need to serve in nvfp4, use MoE only for throughput, reap so training doesn’t OOM, dig into the vllm kernel graph itself to not underutilize my hardware from poor flashinfer/cutlass selections for my rtx pro 6000 sm120 architecture - Might do online distillation from glm 5.2 but for now taking it one step at a time - Slept for a bit then woke up and showered - Fixed an issue with SGLang tensor parallel deadlock on GLM 5.2 architecture with MTP enabled - GLM 5.2 inference is 2-3x faster than coding plans and running on amd boxes - Spent time with family - Hung out with some friends - Recorded some yoctogpt lectures with the revamped notebook (high taste btw) - Setting up dflash training for GLM 5.2
0:30

Timelapse #156 (36 hrs) - Worked with the tiny corp on getting GLM 5.2 running on 8xMI300X (sglang won here) - Launched KernelBench-Mega and updated Kernelbench-Hard with h100 and b200 sweeps - Took care of boring business stuff - Did some training sweeps for specialized technical vocab audio model - Some bugs with putting kernelbench-mega and hard on cloud instances so had to do some reruns. Learned a lot though - Setting up my own local rl infra and profiling concurrency 128 rollouts with vllm. Became clear to me that I need to serve in nvfp4, use MoE only for throughput, reap so training doesn’t OOM, dig into the vllm kernel graph itself to not underutilize my hardware from poor flashinfer/cutlass selections for my rtx pro 6000 sm120 architecture - Might do online distillation from glm 5.2 but for now taking it one step at a time - Slept for a bit then woke up and showered - Fixed an issue with SGLang tensor parallel deadlock on GLM 5.2 architecture with MTP enabled - GLM 5.2 inference is 2-3x faster than coding plans and running on amd boxes - Spent time with family - Hung out with some friends - Recorded some yoctogpt lectures with the revamped notebook (high taste btw) - Setting up dflash training for GLM 5.2

Elliot Arledge

45,447 Aufrufe • vor 4 Monaten

DigitalOcean’s model synthesis tool is now available in public preview. 🆕 A new server-side tool: run multiple models on one inference request and get a single synthesized response, no custom orchestration. Just add it to your existing inference request.
3:22

DigitalOcean’s model synthesis tool is now available in public preview. 🆕 A new server-side tool: run multiple models on one inference request and get a single synthesized response, no custom orchestration. Just add it to your existing inference request.

DigitalOcean

2,192,331 Aufrufe • vor 2 Monaten

GLM 4.7 is one of the strongest open-source coding models available—but most developers aren't prompting it correctly. We put together 10 rules to help you get the most out of it: - Front-load instructions (it has a strong recency bias) - Use firm language: "must" and "strictly" > soft suggestions - Break complex tasks into smaller steps - Disable reasoning for simple tasks, enable it for hard ones - Use critic agents for code review, QA, and validation - Pair it with a frontier model for the hardest 10% of workloads - and more… GLM 4.7 hits 96% on Tau² Bench and 86% on GPQA Diamond. At 1,500 tokens/sec on Cerebras, it's 20x faster than closed-source alternatives on GPUs.
5:30

GLM 4.7 is one of the strongest open-source coding models available—but most developers aren't prompting it correctly. We put together 10 rules to help you get the most out of it: - Front-load instructions (it has a strong recency bias) - Use firm language: "must" and "strictly" > soft suggestions - Break complex tasks into smaller steps - Disable reasoning for simple tasks, enable it for hard ones - Use critic agents for code review, QA, and validation - Pair it with a frontier model for the hardest 10% of workloads - and more… GLM 4.7 hits 96% on Tau² Bench and 86% on GPQA Diamond. At 1,500 tokens/sec on Cerebras, it's 20x faster than closed-source alternatives on GPUs.

Cerebras

633,658 Aufrufe • vor 8 Monaten

oMLX 0.7.1.dev1 is out! This dev release adds decision models (Clef, OpenJev), batched prefill for concurrent requests, a refreshed web UI, and faster Qwen3.8, Qwen3.6, GLM-5.3-Flash and DeepSeek V4.1. Speedups in this release - Qwen3.6-35B-A3B decode on M5 Max: 156 -> 205 tok/s (+31%). - GLM-5.3-Flash decode on M3 Ultra: 31.3 -> 40.6 tok/s (+30%). The fused decode kernels now run on M3 and M4, not only M5. - Qwen3.8-Flash-Next with expert offload, 32K prefill on M3 Ultra: 396 -> 636 tok/s (+61%). - Qwen3.8-Flash-Next, 8 prompts arriving 0.4 s apart on M5 Max: time to first token 7.41 -> 5.75 s (-22%) with batched prefill. What's new - Decision models. Clef, Clef-Flash and OpenJev answer typed questions with a probability for each option through the new /v1/systemone endpoint. - Batched prefill. Prompts that prefill at the same time now share one forward pass, so MoE models read their expert weights once. - A refreshed web UI with a log table, a settings rail and a customizable dashboard. - Native video input for Qwen3.5, 3.6 and 3.8. - More accurate oQe quantization, with 11-15% lower KL to the original model. - Fewer memory guard rejections in long sessions. Most of the speedups came from contributors again. Thanks to everyone who sent PRs and tested the dev builds!
0:33

oMLX 0.7.1.dev1 is out! This dev release adds decision models (Clef, OpenJev), batched prefill for concurrent requests, a refreshed web UI, and faster Qwen3.8, Qwen3.6, GLM-5.3-Flash and DeepSeek V4.1. Speedups in this release - Qwen3.6-35B-A3B decode on M5 Max: 156 -> 205 tok/s (+31%). - GLM-5.3-Flash decode on M3 Ultra: 31.3 -> 40.6 tok/s (+30%). The fused decode kernels now run on M3 and M4, not only M5. - Qwen3.8-Flash-Next with expert offload, 32K prefill on M3 Ultra: 396 -> 636 tok/s (+61%). - Qwen3.8-Flash-Next, 8 prompts arriving 0.4 s apart on M5 Max: time to first token 7.41 -> 5.75 s (-22%) with batched prefill. What's new - Decision models. Clef, Clef-Flash and OpenJev answer typed questions with a probability for each option through the new /v1/systemone endpoint. - Batched prefill. Prompts that prefill at the same time now share one forward pass, so MoE models read their expert weights once. - A refreshed web UI with a log table, a settings rail and a customizable dashboard. - Native video input for Qwen3.5, 3.6 and 3.8. - More accurate oQe quantization, with 11-15% lower KL to the original model. - Fewer memory guard rejections in long sessions. Most of the speedups came from contributors again. Thanks to everyone who sent PRs and tested the dev builds!

Jun Kim

24,706 Aufrufe • vor 10 Tagen

🚨 ALIBABA JUST OPEN-SOURCED THE MOST POWERFUL MODEL IN THE WORLD - Qwen3.5 plus And it changes the AI race forever. This is not another “bigger model” release. This is a complete architectural reset. 397B parameters → only 17B active at inference, Qwen3.5 plus delivers performance on par with GPT-5.2 and Gemini-3-Pro. A new Hybrid Sparse MoE + Linear Attention architecture + Multi-Token Prediction = near-instant long-context responses and massive speed gains. Built as a true native multimodal model, it doesn’t just chat it builds. With a single prompt: “Use the React framework to create a visually impressive Mapbox demo featuring locations in Beijing and Shanghai” it generates a live interactive 3D map UI, structured components, real layouts, and production-ready frontend logic. This is real agentic, multimodal coding not a static code block. MMLU-Pro: 87.8 GPQA: 88.4 IFBench: 76.5 ⚡ up to 8× faster inference And it’s fully open-sourced under Apache-2.0. No locked APIs. No restricted tiers. 201 languages supported bringing frontier AI to developers, startups, and researchers worldwide. The race is no longer about the biggest model. It’s about the most usable intelligence. Open source just entered the top tier.
1:50

🚨 ALIBABA JUST OPEN-SOURCED THE MOST POWERFUL MODEL IN THE WORLD - Qwen3.5 plus And it changes the AI race forever. This is not another “bigger model” release. This is a complete architectural reset. 397B parameters → only 17B active at inference, Qwen3.5 plus delivers performance on par with GPT-5.2 and Gemini-3-Pro. A new Hybrid Sparse MoE + Linear Attention architecture + Multi-Token Prediction = near-instant long-context responses and massive speed gains. Built as a true native multimodal model, it doesn’t just chat it builds. With a single prompt: “Use the React framework to create a visually impressive Mapbox demo featuring locations in Beijing and Shanghai” it generates a live interactive 3D map UI, structured components, real layouts, and production-ready frontend logic. This is real agentic, multimodal coding not a static code block. MMLU-Pro: 87.8 GPQA: 88.4 IFBench: 76.5 ⚡ up to 8× faster inference And it’s fully open-sourced under Apache-2.0. No locked APIs. No restricted tiers. 201 languages supported bringing frontier AI to developers, startups, and researchers worldwide. The race is no longer about the biggest model. It’s about the most usable intelligence. Open source just entered the top tier.

RAVI KUMAR SAHU

380,400 Aufrufe • vor 8 Monaten

Today, we're introducing not one but TWO new models, striking the balance between efficiency and quality to enable you to build production AI agents. — Gemini 3.6 Flash: Addresses efficiency feedback we received from Gemini 3.5 Flash with upgrades in coding, knowledge work, and multimodal tasks faster, more accurately, and with substantially fewer tokens per task — Gemini 3.5 Flash-Lite: Our fastest, most cost-effective 3.5-class model yet built for agentic workflows, hitting ~350 output tokens/sec with improved coding and overall quality Start building with these today via the Gemini API in Google AI Studio or try them out in the Google Gemini
0:42

Today, we're introducing not one but TWO new models, striking the balance between efficiency and quality to enable you to build production AI agents. — Gemini 3.6 Flash: Addresses efficiency feedback we received from Gemini 3.5 Flash with upgrades in coding, knowledge work, and multimodal tasks faster, more accurately, and with substantially fewer tokens per task — Gemini 3.5 Flash-Lite: Our fastest, most cost-effective 3.5-class model yet built for agentic workflows, hitting ~350 output tokens/sec with improved coding and overall quality Start building with these today via the Gemini API in Google AI Studio or try them out in the Google Gemini

Google AI

439,464 Aufrufe • vor 3 Monaten