# generation Published articles for generation. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Micron Shows off 512GB DDR5 RDIMM: 12TB per Dual-Socket Server at 9,200 MT/s, Volume Production in 2H 2027 DevFeed: [Micron Shows off 512GB DDR5 RDIMM: 12TB per Dual-Socket Server at 9,200 MT/s, Volume Production in 2H 2027]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-09-15T20:18:17Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [ddr5](), [servers](), [intel](), [cpu]() Tags: [capacity](), [ddr5](), [enterprise](), [generation](), [intel](), [memory](), [modules](), [performance](), [production](), [release](), [server](), [speed](), [volume]() ### AI overview Micron demonstrated a 512GB DDR5 RDIMM rated for up to 9,200 MT/s. The module can provide 12TB of memory in a 24-slot dual-socket server, with volume production scheduled for the second half of 2027. AMD and Intel are validating it for next-generation server platforms. ### Source excerpt Micron has demonstrated a 512GB DDR5 RDIMM running on multiple server platforms, which it calls the world's first module at that capacity, and says AMD and Intel are both validating it for their next-generation server platforms. The module is rated for speeds up to 9,200 MT/s, and in a 24-slot dual-socket server it puts 12TB The post Micron Shows off 512GB DDR5 RDIMM: 12TB per Dual-Socket Server at 9,200 MT/s, Volume Production in 2H 2027 appeared first on StorageReview.com. ## America is building datacenters faster than the grid can power them DevFeed: [America is building datacenters faster than the grid can power them]() Original publisher: [Read original article]() Author: Dan Robinson Published: 2026-09-15T16:37:00Z Content type: news Language: en Sources: [www.theregister.com - Articles]() Topics: [datacenter](), [systems]() Tags: [ai-and-ml](), [datacenters](), [energy](), [generation](), [on-prem](), [report]() ### AI overview Meeting expected energy consumption through 2030 will require $110 billion in new generation resources as America builds datacenters faster than the power grid can support. ### Source excerpt Meeting expected energy consumption through 2030 will require $110 billion in new generation resources ## Using Exact-Match Response Caching to Reduce LLM Costs DevFeed: [Using Exact-Match Response Caching to Reduce LLM Costs]() Original publisher: [Read original article]() Author: Abhilash Rao Mesala Published: 2026-09-14T11:00:00Z Content type: article Language: en Sources: [The New Stack]() Topics: [Caching](), [Large Language Model]() Tags: [ai-engineering](), [ai-infrastructure](), [caching](), [contributed](), [cost](), [finops](), [generation](), [hash](), [llm](), [token]() ### AI overview The article explains how to reduce LLM costs by fingerprinting requests, context, model settings, and underlying data to create exact-match cache keys. Valid cached responses can be reused without calling the model. It distinguishes response caching from provider prompt caching, where only eligible prompt computation is reused. ### Source excerpt An LLM can answer the same question a thousand times and charge you each time. Before paying for another answer, The post Why an old caching trick is your secret to lower LLM costs appeared first on The New Stack. ## HeyGen x Google Cloud: Bringing Avatar IV to TPUs DevFeed: [HeyGen x Google Cloud: Bringing Avatar IV to TPUs]() Original publisher: [Read original article]() Published: 2026-09-12T11:04:33.891311Z Content type: article Language: en Sources: [Google Developers Blog]() Topics: [Google](), [Cloud](), [Artificial Intelligence](), [Optimization](), [AI Infrastructure](), [Transformer](), [PyTorch](), [Compiler](), [Streaming](), [real-time](), [Hardware](), [API]() Tags: [ai](), [ai-infrastructure](), [api](), [cloud](), [code](), [compiler](), [engineering](), [generation](), [google](), [hardware](), [model](), [optimization](), [performance](), [pytorch](), [real-time](), [streaming](), [time](), [tpu](), [transformers](), [video]() ### AI overview HeyGen and Google Cloud describe porting the 18B+ parameter Avatar IV talking-head video generation pipeline to an eight-chip Trillium TPU host. Using torchax, JAX, XLA, FSDP sharding, Ulysses sequence parallelism, and custom Pallas kernels, the team improved performance by 1.86x for real-time chunked streaming while preserving output quality through strict quality gates. ### Source excerpt HeyGen ported their 18B+ parameter Avatar IV video generation model to Google Cloud's Trillium (v6e) TPUs via torchax and XLA, utilizing FSDP and Ulysses sequence parallelism across an eight-chip mesh. To achieve a 1.86x speedup for real-time streaming, the engineering team pipelined exposed all-to-all collectives, aligned sparse attention block sizes to eliminate mask padding, and bypassed softmax serial dependencies using a precomputed Cauchy-Schwarz upper bound. These custom Pallas kernel and compiler optimizations were deployed only after passing rigorous two-tier quality gates to guarantee byte-identical or mathematically equivalent pixel outputs. ## SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign DevFeed: [SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign]() Original publisher: [Read original article]() Published: 2026-09-11T00:00:00Z Content type: article Language: en Sources: [Apple Machine Learning Research]() Topics: [multimodal](), [Transformer](), [datasets]() Tags: [benchmarks](), [drug-discovery](), [generation](), [model](), [multimodal](), [training]() ### AI overview SimpleDesign is a single-stage, end-to-end multimodal generative model for jointly designing protein sequences and three-dimensional structures. It uses Transformer-based multimodal backbones, trains directly in data space on more than 2 million sequence-structure pairs, and achieves competitive results on co-design and unconditional generation benchmarks. ### Source excerpt Proteins are fundamental to biological processes, with their function determined by the complex interplay between the amino acid sequence and the three-dimensional structure. Developing generative models capable of understanding this intrinsically multi-modal relationship is crucial for fields like drug discovery and protein engineering. Existing models often rely on a multi-stage training process where autoencoders that tokenize data into latent representations are trained in a first stage. Secondly, a generative model is trained on the latent representation of the autoencoder(s), i.e... ## ToolGrad: Efficient tool-use dataset generation with textual "gradients" DevFeed: [ToolGrad: Efficient tool-use dataset generation with textual "gradients"]() Original publisher: [Read original article]() Published: 2026-09-10T22:50:22Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [dataset](), [Large Language Model](), [LLM Techniques](), [Machine learning](), [Fine-tuning](), [LLM evaluation / benchmarking]() Tags: [agent](), [cost](), [datasets](), [evaluation](), [fine-tuning](), [generation](), [llm](), [machine-intelligence](), [machine-learning](), [natural-language-processing](), [prompt-engineering](), [tool]() ### AI overview ToolGrad generates tool-use chains before deriving corresponding user queries, aiming to create complex training data for LLM tool use more efficiently and at lower cost than exploration-based approaches. ### Source excerpt Machine Intelligence ## GPT Images 2.5 promises edits that leave the rest of your image alone DevFeed: [GPT Images 2.5 promises edits that leave the rest of your image alone]() Original publisher: [Read original article]() Author: Meredith Shubel Published: 2026-09-10T19:46:32Z Content type: news Language: en Sources: [The New Stack]() Topics: [Artificial Intelligence](), [Latency]() Tags: [ai](), [applications](), [cost](), [generation](), [gpt](), [images](), [latency](), [model](), [models](), [openai](), [production](), [prototyping](), [search](), [workflows]() ### AI overview OpenAI's GPT Images 2.5 introduces two image-editing models: Flare, optimized for speed and lower latency, and Sunburst, designed for greater precision and control. Both have the same listed token rates, but OpenAI does not explain their actual token consumption or comparative per-image costs. ### Source excerpt When OpenAI launched GPT Images 2.5 this week, the company promised better results for a common editing task: changing one The post GPT Images 2.5 promises edits that leave the rest of your image alone appeared first on The New Stack. ## Powering the AI era: How wave energy can complement a 24/7 energy mix DevFeed: [Powering the AI era: How wave energy can complement a 24/7 energy mix]() Original publisher: [Read original article]() Author: Elias Habbar-Baylac Published: 2026-09-10T15:20:22Z Content type: article Language: en Sources: [Cisco Blogs]() Topics: [AI Infrastructure](), [data centers](), [datacenter](), [Artificial Intelligence]() Tags: [ai](), [ai-infrastructure](), [chief-sustainability-office](), [cisco-purpose](), [data-center](), [data-centers](), [energy](), [environmental-sustainability](), [generation](), [infrastructure](), [megawatt](), [our-corporate-purpose](), [sustainability]() ### AI overview The article explains how wave energy could complement solar, wind, and batteries in meeting the continuous electricity needs of AI-era data centers. It discusses a modeled 100 MW flat-load data center and energy portfolios evaluated for cost, reliability, emissions, and round-the-clock availability. ### Source excerpt CorPower Ocean, a Cisco Investments portfolio company, has explored how wave energy technology could complement other sources for 24/7 energy needs. ## Temporally stable generative illumination with a one-step diffusion model DevFeed: [Temporally stable generative illumination with a one-step diffusion model]() Original publisher: [Read original article]() Author: SungYe Kim; Harish Anand; Alexandr Kuznetsov; Wojciech Uss; Wojciech Kaliński; Rama Harihara Published: 2026-09-09T13:00:00Z Content type: article Language: en Sources: [AMD GPUOpen]() Topics: [real-time rendering](), [VAE]() Tags: [arr-group](), [article-release](), [diffusion](), [generation](), [gi](), [inference](), [lighting](), [ml](), [model](), [optimization](), [quality](), [ray-tracing](), [raytracing](), [real-time](), [real-time-rendering](), [research](), [white-paper]() ### AI overview The article presents a single-step latent diffusion method for real-time global illumination. It conditions image generation on scene signals and lighting hints, and uses a Temporal VAE decoder with motion-vector reprojection to improve temporal stability and reduce flicker. ### Source excerpt A generative method for real-time global illumination using a single-step latent diffusion model, delivering stable, high-quality lighting without costly iterative processing. ## Introducing ChatGPT Images 2.5 DevFeed: [Introducing ChatGPT Images 2.5]() Original publisher: [Read original article]() Published: 2026-09-08T11:30:00Z Content type: release Language: en Sources: [OpenAI News]() Topics: [API]() Tags: [api](), [chatgpt](), [codex](), [feature](), [generate](), [generation](), [gpt](), [images](), [latency](), [mobile](), [models](), [product](), [speed](), [web](), [work](), [workflows]() ### AI overview ChatGPT Images 2.5 is a new image model with sharper details, more precise multi-turn editing, improved reference-photo fidelity, richer textures, and up to 50% lower generation latency than Images 2.0. The release adds creative controls in ChatGPT and introduces GPT-Image-2.5 Flare and Sunburst in the API. ### Source excerpt ChatGPT Images 2.5 helps turn your ideas, sketches, and reference photos into more personalized, polished images that better reflect your ideas. ## GPT Image 2.5 Flare and Sunburst now available on AI Gateway DevFeed: [GPT Image 2.5 Flare and Sunburst now available on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-09-08T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [Artificial Intelligence](), [OpenAI]() Tags: [ai](), [generation](), [gpt](), [image](), [images](), [models](), [openai](), [playground]() ### AI overview OpenAI's GPT Image 2.5 Flare and GPT Image 2.5 Sunburst are now available through AI Gateway for image generation and editing. Flare emphasizes faster generation, while Sunburst prioritizes editing precision; both support text prompts, reference images, detailed instructions, complex layouts, transparent backgrounds, and targeted edits that preserve the rest of an image. ### Source excerpt GPT Image 2.5 Flare and GPT Image 2.5 Sunburst from OpenAI are now available on AI Gateway. Both models accept text prompts and reference images for generation and editing. They produce natural lighting and textures, follow detailed visual instructions, handle complex layouts and transparent backgrounds, and make targeted edits while preserving the rest of an image. Choose Flare for faster generation and Sunburst when editing precision matters most. For Flare, use the model ID openai/gpt-image-2.5-flare: For Sunburst, use openai/gpt-image-2.5-sunburst. Pass a reference image with the instruction: Try Flare or Sunburst in the model playground, or view all image models available on AI Gateway. Read more ## Test what you ship: MSTest and Native AOT DevFeed: [Test what you ship: MSTest and Native AOT]() Original publisher: [Read original article]() Author: Amaury Levé Published: 2026-09-03T18:00:00Z Content type: article Language: en Sources: [.NET Blog]() Topics: [native aot](), [trimming](), [.NET](), [Code generation](), [Deployment]() Tags: [deployment](), [generation](), [mstest](), [native-aot](), [net](), [source-generators](), [testing](), [trimming]() ### AI overview MSTest 4.4 uses source generation to let test projects run as Native AOT executables, helping teams test under the same trimming and deployment constraints as their applications. The article explains that this can reveal application issues such as reflection-dependent serialization that managed test runs may miss. ### Source excerpt MSTest source generation lets test projects use the same Native AOT and trimming deployment model as the applications they validate, while reducing reflection on the test execution path. The post Test what you ship: MSTest and Native AOT appeared first on .NET Blog. ## Scaling llms.txt with a Hierarchy of Files DevFeed: [Scaling llms.txt with a Hierarchy of Files]() Original publisher: [Read original article]() Author: Kyan Yang Published: 2026-09-01T00:00:00Z Content type: article Language: en Sources: [Mintlify Blog]() Topics: [Documentation](), [navigation](), [Website]() Tags: [agents](), [context-window](), [documentation](), [engineering](), [files](), [generation](), [navigation]() ### AI overview Mintlify rebuilt llms.txt generation as a hierarchy of files to keep large documentation sites within the 100,000-character limit while preserving access to pages. In tests across 200 tasks on ten documentation sites, the improved version completed tasks 52.1% faster and used 44.8% fewer tokens. ### Source excerpt Large documentation sites were hitting Mintlify's 100,000-character llms.txt limit and omitting pages. We rebuilt generation as a hierarchy of files so agents can reach every page without loading the entire index. ## MiniMax H3 and H3 Max are 50% off on AI Gateway DevFeed: [MiniMax H3 and H3 Max are 50% off on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-08-30T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [Artificial Intelligence](), [API](), [API keys](), [dashboards](), [browser]() Tags: [ai](), [ai-gateway](), [api](), [audio](), [browser](), [generate](), [generation](), [image](), [model](), [models](), [pricing]() ### AI overview MiniMax H3 and H3 Max receive a 50% discount on Vercel AI Gateway from August 30 through September 13. H3 supports 2K video generation from text, images, video, and audio inputs, while H3 Max offers faster 480p and 768p rendering from text or a starting image. Existing model IDs remain unchanged, so no code changes are required. ### Source excerpt MiniMax H3 and H3 Max are 50% off on AI Gateway from August 30 through September 13, in partnership with MiniMax. The discount covers requests billed through AI Gateway, at every duration and in every aspect ratio the model supports. H3 generates 2K video from a text prompt, a starting image, a pair of first and last frames, or reference images, video, and audio. H3 Max trades resolution for speed: it renders faster at 480p and 768p, and it takes a text prompt or a starting image. The model IDs (minimax/minimax-h3 and minimax/minimax-h3-max) are unchanged, so requests you already send pick up the discounted rate with no code change: Renders take minutes, so poll runs the generation as a background job and makes short status requests until it lands, rather than holding one long request open. See asynchronous generation for the webhook and start-and-status routes. Get started Create an API key in the AI Gateway section of your dashboard, or generate a clip in the browser first from the model playground. Current rates for every model are on the pricing page. You can view all video models available on AI Gateway. Read more ## AMD FSR plugin updated for Unreal Engine 5.8 DevFeed: [AMD FSR plugin updated for Unreal Engine 5.8]() Original publisher: [Read original article]() Author: Joe Rozek; Alexander Blake-Davies Published: 2026-08-27T12:00:00Z Content type: release Language: en Sources: [AMD GPUOpen]() Topics: [Unreal Engine](), [releases](), [GPU](), [Machine learning](), [SDKs]() Tags: [amd-fsr-3](), [amd-fsr-4](), [amd-fsr-frame-generation](), [amd-fsr-framegeneration](), [amd-fsr-upscaling](), [announcement](), [antilag](), [fsr-3](), [fsr-4](), [game-development](), [generation](), [gpu](), [ml](), [news](), [performance](), [plugin](), [product-blogs](), [product-release](), [redstone](), [release](), [super-resolution](), [technical-articles](), [unreal](), [unreal-engine]() ### AI overview The AMD FSR Unreal Engine plugin has been updated for Unreal Engine 5.8. It adds FSR Redstone SDK 2.3 updates, FSR Upscaling 4.1.1 support for AMD Radeon RX 7000 Series GPUs, and FSR Frame Generation 4.0.1 improvements. ### Source excerpt The updated AMD FSR™ Unreal® Engine plugin brings ML-powered upscaling and frame generation to Unreal Engine 5.8, now extending FSR Upscaling support to AMD Radeon RX 7000 Series GPUs. ## How Speculative Decoding Can Make LLM Generation 2-3 Times Faster DevFeed: [How Speculative Decoding Can Make LLM Generation 2-3 Times Faster]() Original publisher: [Read original article]() Author: ByteByteGo Published: 2026-08-26T15:30:34Z Content type: tutorial Language: en Sources: [ByteByteGo]() Topics: [Large Language Model](), [text-generation](), [GPU]() Tags: [article](), [data-center](), [decoding](), [generation](), [gpu](), [how-to](), [llms](), [tokens]() ### AI overview This tutorial explains speculative decoding, in which a smaller model proposes candidate tokens and a larger model evaluates them in a single forward pass. It covers autoregressive generation, GPU utilization, candidate acceptance and rejection, output-quality preservation, acceptance rates, draft sources, and when the technique may stop helping. ### Source excerpt In this article, we will look at how speculative decoding works. ## 【etcd】MVCC 数据模型：Revision、keyIndex 与 generation DevFeed: [【etcd】MVCC 数据模型：Revision、keyIndex 与 generation]() Original publisher: [Read original article]() Author: Liao Tonglang Published: 2026-08-26T00:00:00Z Content type: tutorial Language: zh Sources: [土法炼钢 - 系统与基础设施]() Topics: [etcd](), [Raft](), [Kubernetes]() Tags: [compaction](), [distributed](), [etcd](), [generation](), [keyindex](), [kubernetes](), [mvcc](), [raft](), [revision](), [treeindex](), [v3](), [v3-5](), [v3-5-33](), [watch]() ### AI overview This article analyzes etcd v3.5.33's MVCC data model through three concepts: globally ordered revisions, keyIndex generations that track key lifecycles, and the division of responsibilities between the in-memory treeIndex and the bbolt backend. It explains how these structures support historical Watch replay, transaction comparisons, compaction, and Kubernetes resourceVersion semantics, while distinguishing etcd revisions from Raft indexes. ### Source excerpt 拆解 etcd v3.5.33 的 Revision (main, sub) 全序、keyIndex/generation 生命周期与 treeIndex 分工；简要对照 v2 平面键空间，交代 MVCC 与 Watch/compaction 的语义基础。 ## Luce: Relightable Gaussians for 3D Asset Generation DevFeed: [Luce: Relightable Gaussians for 3D Asset Generation]() Original publisher: [Read original article]() Published: 2026-08-26T00:00:00Z Content type: article Language: en Sources: [Apple Machine Learning Research]() Topics: [Artificial Intelligence](), [Benchmark]() Tags: [3d](), [ai](), [benchmark](), [computer-vision](), [generation](), [images](), [mesh](), [models](), [multimodal](), [techniques](), [vision]() ### AI overview Luce is a multimodal 3D representation for generating relightable assets from a single image. It combines geometry with physically based materials in a voxelized Gaussian cloud, compresses them into a material-aware latent space, and generates relightable PBR Gaussians and optional textured meshes. On Toys4K, it reports a 28% FID improvement over the strongest baseline and improves alignment on an AI-generated image benchmark. ### Source excerpt High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using dedicated Gaussian primitives for each modality. A variational autoencoder compresses this representation into a unified material-aware latent space. A... ## Leading Publishers Bring Blockbuster PC Games and Technology to NVIDIA RTX Spark DevFeed: [Leading Publishers Bring Blockbuster PC Games and Technology to NVIDIA RTX Spark]() Original publisher: [Read original article]() Author: Alexander Mejia Published: 2026-08-25T15:30:48Z Content type: news Language: en Sources: [NVIDIA Blog]() Topics: [Nvidia](), [Artificial Intelligence]() Tags: [conference](), [developers](), [game-development](), [gamescom](), [gaming](), [geforce](), [geforce-now](), [generation](), [launch](), [model](), [nvidia](), [nvidia-rtx](), [pc](), [performance](), [platform](), [ray-tracing](), [rtx-spark](), [technology]() ### AI overview NVIDIA is bringing RTX Spark gaming PCs to Gamescom with support from major publishers including Electronic Arts, Embark and Ubisoft. The platform combines NVIDIA RTX technologies, Windows gaming, anti-cheat support and AI-enabled capabilities, with a launch planned for this fall. ### Source excerpt NVIDIA is bringing the next wave of RTX gaming to the Gamescom conference running this week in Cologne, Germany, with support for new games, anti-cheat technologies and increased visual quality. Electronic Arts, Embark and Ubisoft are among the latest game publishers and developers bringing their blockbuster titles to NVIDIA RTX Spark ahead of its launch [...] ## AI Gateway now supports asynchronous video generation DevFeed: [AI Gateway now supports asynchronous video generation]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-08-25T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [vercel ai sdk](), [SDKs](), [web applications]() Tags: [ai](), [ai-gateway](), [event](), [generation](), [http](), [image-to-video](), [text-to-video](), [video](), [workflow]() ### AI overview Vercel AI Gateway adds asynchronous video generation through workflow webhooks, direct webhooks, polling, or later retrieval, avoiding long-lived requests that may time out. ### Source excerpt Video generation on AI Gateway can now run asynchronously. By default, generateVideo keeps one HTTP request to AI Gateway open until the result is ready. Because video generation can take seconds or minutes, that request can exceed request timeouts. With asynchronous generation, your application can receive a webhook, poll for completion, or start a generation and retrieve the result in a later request. Choose an option based on whether your process can keep running and whether your application can receive webhooks: Existing generateVideo calls continue to work as before. All four options support text-to-video, image-to-video, reference-to-video, and other video inputs. Upgrade the SDK Install the latest versions of the AI SDK and AI Gateway provider: Use asynchronous video generationWait for completion in a Workflow An easy way to consume the completion webhook is a Workflow SDK. The workflow creates its own webhook URL, passes it to startVideo, and suspends until AI Gateway delivers the completion event. Install the Workflow SDK alongside the AI SDK: While the video renders, the workflow run is suspended and resumes when AI Gateway delivers the terminal event. Use a webhook with generateVideo Pass webhook to generateVideo to wait for a completion event without polling. AI Gateway sends an event when the job completes or fails. The SDK waits for that event, fetches the generated videos, and resolves the original generateVideo call. The calling process and webhook handler need a shared token and store so the delivery can be matched to the correct generation. generateVideo does not expose the signing secret for this job. See the webhook verification documentation for the complete receiver pattern. Both the polling and webhook options for generateVideo return result.videos as GeneratedFile objects. The SDK downloads provider-hosted videos, making uint8Array, base64, and mediaType available in either job. Poll with generateVideo Add poll to an existing generateVideo ca ## Wan 3.0 now available on AI Gateway DevFeed: [Wan 3.0 now available on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-08-25T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [Artificial Intelligence]() Tags: [ai](), [ai-gateway](), [generation](), [image-to-video](), [model](), [models](), [text-to-video](), [video]() ### AI overview Wan 3.0 from Alibaba is now available on Vercel's AI Gateway as alibaba/wan-v3.0-video. The model supports text-to-video, image-to-video, first- and last-frame conditioning, and reference-based generation using images, video, and audio. It can generate clips up to 30 seconds at 30 fps in resolutions up to 1080p with synchronized audio, and supports asynchronous generation through webhooks. ### Source excerpt Wan 3.0 from Alibaba is now available on AI Gateway as alibaba/wan-v3.0-video. Wan 3.0 combines text-to-video, image-to-video, first- and last-frame conditioning, and reference-based generation in one model. References can include images, video, and audio. It generates clips up to 30 seconds at 30 fps in 480p, 720p, or 1080p, with synchronized audio. Previously, Wan 2.7 required separate -t2v and -r2v model IDs and was limited to 15-second clips at 24 fps. Generate a video Wan 3.0 supports asynchronous generation, so no HTTP request needs to remain open for the entire render. Pass a webhook to receive an event when the generation finishes: Learn about asynchronous generation options in the docs, including how to verify webhook deliveries. Add references Pass image, video, or audio references through inputReferences, including the source and media type for each one. Images accept hosted URLs or base64. Video and audio references require hosted URLs. First- and last-frame conditioning accepts one image for each frame and can't be combined with other references. Try Wan 3.0 in the model playground, or browse all video models. Read more ## Configuring compaction thresholds and context windows for coding agents DevFeed: [Configuring compaction thresholds and context windows for coding agents]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-08-14T19:33:33Z Content type: tutorial Language: en Sources: [WorkOS Blog]() Topics: [coding](), [context window](), [long-context]() Tags: [agent](), [coding](), [context-window](), [generation](), [model](), [recovery](), [tokens]() ### AI overview The article explains how to derive a coding agent's effective context window from compaction thresholds and the response runway required to complete generation. It discusses threshold behavior, model metadata, route limits, generation clamping, overflow detection, and recovery. ### Source excerpt Derive your coding agent's effective context window from two numbers: the compaction threshold you want, and the response runway the model needs to finish. ## Introducing CARE-X: Towards Clinically Useful Radiology VLMs with Auxiliary Supervision, Reward-Aligned Learning, and Tool-Augmented Measurement DevFeed: [Introducing CARE-X: Towards Clinically Useful Radiology VLMs with Auxiliary Supervision, Reward-Aligned Learning, and Tool-Augmented Measurement]() Original publisher: [Read original article]() Author: Mercy Ranjit, Nikhilesh E, Dr. Abhyuday Kumara Swamy, Tanuja Ganu Published: 2026-08-11T16:00:00Z Content type: article Language: en Sources: [Microsoft Research]() Topics: [vlm](), [Artificial Intelligence](), [Reinforcement learning](), [data]() Tags: [ai](), [data](), [generation](), [reasoning](), [reinforcement-learning](), [research](), [research-blog](), [tools](), [vlms]() ### AI overview CARE-X is a research chest X-ray vision-language model that combines free-text report generation, structured diagnostic prediction, and reinforcement learning for multi-task clinical interpretation. The article also describes a separate experiment using deterministic measurement tools with Qwen3-VL-4B-Instruct and reports validation on real-world Indian clinical data, while emphasizing that CARE-X is not approved for clinical use. ### Source excerpt Radiology AI is evolving beyond report generation. CARE-X explores a unified approach that combines flexible reasoning, calibrated predictions, and measurement-based tools for chest X-ray interpretation. The post Introducing CARE-X: Towards Clinically Useful Radiology VLMs with Auxiliary Supervision, Reward-Aligned Learning, and Tool-Augmented Measurement appeared first on Microsoft Research. ## Generate Trajectories, Reasoning Traces, and Auto-Labels with NVIDIA Alpamayo 2 Super DevFeed: [Generate Trajectories, Reasoning Traces, and Auto-Labels with NVIDIA Alpamayo 2 Super]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-08-04T15:00:00Z Content type: tutorial Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [foundation-models](), [Cosmos](), [model-deployment](), [Reinforcement learning](), [Inference](), [GitHub](), [hugging face](), [Open Source Models & Datasets]() Tags: [accelerate](), [automotive-transportation](), [autonomous](), [benchmarks](), [cosmos](), [customization](), [data](), [developer-tools-techniques](), [developers](), [development](), [diffusion](), [drive](), [driving](), [evaluation](), [featured](), [fine-tuning](), [generate](), [generation](), [github](), [hugging-face](), [inference](), [learning](), [models](), [nvidia](), [reasoning](), [robot-navigation](), [robotics](), [simulation-modeling-design](), [top-stories]() ### AI overview NVIDIA Alpamayo 2 Super is an open 34-billion-parameter reasoning vision-language-action model for autonomous vehicle development. It combines NVIDIA Cosmos 3 Super Reasoner with a diffusion-based Action Expert to generate trajectories, reasoning traces, meta-actions, scene answers, and auto-labels across development workflows. ### Source excerpt Autonomous vehicle (AV) development often relies on separate models for trajectory generation, high-level intent prediction, scene understanding, and data... [Next page]()