Which AI model actually fits the job? You won't know until you experiment. The Alibaba Cloud Token Plan makes it easier to try, test and compare, with: · One credit pool across Qwen, Wan, HappyHorse, DeepSeek and GLM · Text, audio, image and video capabilities · Clearer visibility into your AI usage And it’s easy to start, with your first month priced at just $4. Explore the Token Plans: #AlibabaCloud# #TokenPlan# #Qwen# #Wan# #HappyHorse# #DeepSeek# #GLM# #MultiModelAI# #GenerativeAI# #AIWorkflow#

Check out How an Automated Credit Decisioning System Works at @MultimodalAI 😱(Step by Step) A typical automated credit decisioning workflow includes: 1. Case setup and intake Applications and documents are ingested, validated, and normalized automatically. Missing data is flagged early, reducing back-and-forth and manual entry. 2. Intelligent diligence Instead of reviewing everything, the system highlights only the areas that matter; based on policy thresholds, risk signals, and data gaps. 3. Micro-decisioning against policy Identity checks, consistency checks, eligibility rules, and policy constraints are evaluated step by step, not as a single opaque decision. 4. Final decision and routing Applications are approved, declined, or conditionally approved. Edge cases are routed to human reviewers with clear context. 5. Confidence scoring and justification Every decision is paired with a confidence score and explanation, supporting auditability and regulatory review. 6. Reporting and documentation Full decision trails, exceptions, and outcomes are automatically documented for compliance and transparency. The result: faster decisions, lower operational risk, and consistent policy enforcement without sacrificing human oversight. #productvideo# #AutomatedCreditDecisioningSystem#

🦢 TL;DR: nails the timing of listening, speaking and interrupting, but task completion still lags. Tencent unveils an omni interaction model unifying speech, video and agentic execution. Title: Omni Interaction Agent Technical Report (Gander) URL: 📌 Highlights 🧠 Cerebellum-Brain design: real-time control layer + Claude/Codex-powered reasoning layer 🎙 Streams audio/video in 1-second chunks, no external VAD needed ✅ 100% turn-taking accuracy vs GPT-Realtime's 96% ⚠️ Task completion (Pass@1) only 0.400, below GPT-Realtime's 0.600 🔀 Feeding transcribed text straight to the Brain lifts Pass@1 to 0.520 🌐 Fusing audio+video boosts Daily-Omni by +19.13pt over single-modality 📦 Weights, training code and eval harness planned for release The honest takeaway: nailing conversational timing and nailing task accuracy are still two separate problems. #MultimodalAI# #VoiceAgents#

Why manage separate AI tools for the script, the image, and the video? The Alibaba Cloud Token Plan gives you one shared credit pool across supported models and tools, with visibility into usage and access to newer models like Qwen3.8-Max-Preview, HappyHorse1.1, DeepSeek V4, and GLM-5.2. One plan for every modality, to build more, spend less. Get started from just $4 in your first month. Explore the Token Plans: #AlibabaCloud# #TokenPlan# #Qwen# #Wan# #HappyHorse# #GenerativeAI# #AIContent# #MultimodalAI#

TL;DR A new suite trains and evaluates "native visual reasoning," where visual generation itself is the medium of reasoning, using large-scale data and verifiable rewards. VLM-judge scores swung by up to 92.8% on the same video. Title: VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoning URL: Key points 📊 300 tasks, 1.25M training instances, ~3.47M images and 1.3M videos in a large-scale dataset 🎯 A deterministic scorer using classical CV (HSV color segmentation, OCR) hits 0.60+ human agreement, beating GPT-5.5 and Gemini-3.1-Pro 🌀 Coefficients-Preserving Sampling (CPS) keeps predicted and fresh noise coefficients balanced, stabilizing RL exploration 🚀 Transfer to V-ReasonBench jumps from 10.21 to 38.22 (+28.01 pts), the largest gain across seven external benchmarks 🔬 Counterfactual tests show removing input images drops scores by 90%, confirming visual trajectories matter more than text for reasoning ✅ Verifiable-reward RL beats VLM-reward RL by +7.9% in-domain What stands out is moving visual reasoning away from language dependence toward something trainable and verifiable in its own right. #VisualReasoning# #MultimodalAI#

Video AI has been reasoning over blurred and occluded footage while trusting every frame equally. This work tackles that blind spot. Title: Confidence-Aware Tool Orchestration for Robust Video Understanding URL: ❓ What's the problem? 💡 Video-LLMs implicitly assume every frame is equally reliable (the authors' "Blind Trust Problem"). When footage degrades from motion blur, glare, or occlusion, they fail to notice and lose 15-30 points on real-world benchmarks, while their self-reported confidence barely changes, a silent failure. ❓ How does Robust-TO solve it? 💡 It bakes per-frame trustworthiness into every reasoning stage. First, quality profiling scores blur, brightness, and occlusion to keep only reliable frames; then it decomposes the query into sub-queries routed to tools robust to the dominant corruption, and every tool returns a (result, confidence) pair. ❓ How is confidence used? 💡 Evidence is grouped into high/medium/low tiers. High drives the conclusion, medium is kept only when consistent, low is a fallback only, and any residual uncertainty is stated in the answer. It is trained with GRPO using a confidence-cost reward. ❓ How well does it work? 💡 56.4% average on clean video (+10.6pt over Gemini-2.5-Pro) and 54.3% under corruption (+5.8pt over the strongest open-source Video-R1). It also trims frames from 32 to 20.7, cutting inference time by over 35% while gaining +1.6pt accuracy. #VideoUnderstanding# #MultimodalAI#

Inside MiniMax H3: One DiT Stream for Text, Video, and Stereo Audio Earlier this month, @MiniMax_AI open-sourced H3, a multimodal model that accepts text, images, video, and audio, then jointly generates video with native stereo sound. The interesting part is not just the capability list. It is how H3 represents several modalities inside one diffusion transformer. Zhihu contributor 微卷的大白 analyzed the released code and checkpoints. Since MiniMax had not published the full technical report, some low-level details should be treated as code-based interpretation rather than official specification. 1️⃣ What exactly was open-sourced? H3 supports 4–15 seconds of video at 24 fps, together with 32 kHz stereo audio. The released model has two main checkpoints: 🔹 FL2VA handles text-to-audio-video and generation conditioned on optional first or last frames. 🔹 Ref2VA accepts mixed references, including images, videos, and audio. The open H3-Base path generates at 768p. The full 2K product pipeline relies on In-Context Regeneration, which was not open-sourced when the analysis was written. The complete Contextual Omni Representation processing chain and Native Sparse Attention were also not fully available. This distinction matters when evaluating local results or inference cost. 2️⃣ Every modality enters one packed sequence H3 uses a 50-layer Omni Transformer with a hidden size of 5,376. Text conditions, reference media, noisy video latents, and noisy audio latents are packed into the same attention sequence. They share one set of attention projections and the same SwiGLU feed-forward network. Only the target video and audio rows are updated during Euler denoising. Reference and conditioning rows remain context. At the output, two separate heads predict video and audio velocities from the shared hidden states. H3 is not generating video first and attaching sound afterward. Both modalities evolve inside the same denoising process. 3️⃣ The Token Refiner bridges understanding and generation Before entering the DiT, representations from Qwen3-VL pass through a two-layer Token Refiner. A simple linear projector can only transform each token independently. Self-attention allows every conditioning token to reread and reorganize the complete prompt context. The refiner does not see noisy video or audio latents, and it does not perform denoising. Its job is to convert the understanding model’s output into conditioning that the generative backbone can use. The author also found that short prompts often produced weak results. Rewriting them in the richer style of MiniMax’s official examples substantially improved generation quality. 4️⃣ RoPE creates a shared physical timeline The hardest positional problem is that one attention stream must represent several different structures: 🔹 Text has sequential order. 🔹 Video has time, height, and width. 🔹 Audio has time and stereo-channel identity. H3 solves this with three-axis positional coordinates. Video frames and audio samples are mapped onto a shared physical timeline, while the other axes encode spatial position or audio channel. Audio latents run at 40 Hz, while video runs at 24 fps. H3 therefore advances video time by 5/3 units per source frame so audio and video can periodically align on the same coordinates. Left and right audio channels share the same time coordinate but use different positions on another axis. This preserves synchronization while retaining channel identity. Importantly, packed row order does not define physical time. RoPE coordinates do. 5️⃣ Modulation is huge in parameters, tiny in FLOPs Each DiT block generates shift, scale, and gate parameters for both attention and MLP paths. The parameters are selected according to two signals: 🔹 The diffusion timestep 🔹 Whether the row represents text, video, or audio Across 50 layers, these AdaLN-related projections contain roughly 13 billion parameters, around 39% of the DiT. That sounds computationally expensive, but the projections operate on a small table of unique timesteps and modalities. The resulting parameters are then gathered for each row. For a five-second generation, this part contributes less than 0.002% of forward-pass FLOPs. It is a striking design choice: parameter-heavy conditioning without token-proportional projection cost. 6️⃣ Long video turns attention into the bottleneck After VAE compression and DiT patching, a five-second 768p sample still contains roughly 37,700 effective rows. At ten seconds, that grows to about 73,400. At fifteen seconds, it exceeds 109,000. With full attention: ✅ The sequence grows by about 2.9× from five to fifteen seconds. ✅ Compute per DiT forward grows by roughly 6.1×. ✅ Attention’s share of compute rises from 58.4% to 80.2%. H3 uses a distilled CFG path and executes 49 DiT forwards across its sigma schedule. For a fifteen-second sample, the author estimates aggregate DiT computation at roughly 1.04 exaFLOPs. Since Native Sparse Attention was not included in the initial open release, the public inference path analyzed here still pays the quadratic cost of full attention. That makes sparse attention and fused kernels the clearest opportunities for infrastructure optimization. 🔍 The architectural takeaway H3’s core idea is a shared generative space. Text provides instructions. Reference media supplies context. Video and stereo audio are denoised together. RoPE aligns them in space and physical time, while indexed modulation tells each row how to behave. The model’s biggest strength is therefore not simply “audio-video generation.” It is the attempt to make multiple modalities behave like one coordinated sequence. Its biggest constraint is equally clear: as duration grows, full attention rapidly becomes the dominant cost. 🔗 Full analysis: #MiniMaxH3# #VideoGeneration# #DiffusionTransformer# #MultimodalAI# #GenerativeAI# #AIInfra#

‘This special issue proposes multimodality as an integrated and versatile methodological and interpretative framework for museum research and practice, while showcasing innovative methods as a way of advancing multimodality research…’ Access issue here -