Can video generation models actually reason about the visual world — or do they just make things look plausible? These sound similar but are fundamentally different. Placing objects correctly, respecting physical laws, and solving multi-step tasks with coherent intermediate states: this is what "Visual Grounded Intelligence" means. VGI-Bench quantifies how far current models are from achieving it. Across 27 tasks and 810 instances, the strongest model — Seedance 2.0 — reaches only 51.0% overall. Three failure modes appear repeatedly: physical collapse (unrealistic deformations and object penetrations), rule violations (reaching plausible end states while ignoring constraints), and object inconsistency (losing temporal identity across frames). Open-source models lag far behind commercial ones, with the strongest open-source model at 19.1%. What the denoising analysis reveals is even more telling. Self-correction during intermediate generation steps occurs in fewer than 1% of transitions. Wrong-to-wrong transitions at mid-denoising reach 23.1%, and later denoising steps primarily refine early hypotheses rather than fixing errors. Fine-tuning on 1M synthetic samples improved planning and spatial skills but degraded temporal and physical capabilities — a tradeoff that points to deeper architectural constraints. VGI-Bench makes clear that visual reasoning remains a genuine unsolved frontier for video generation. Title: VGI-Bench: Probing Visual Intelligence in Video Generation Models URL: #VideoGeneration# #VisualReasoning#

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#

GPT-6 Astra really looks like openai went hard on pre-training world modeling, math/abstraction, spatial + visual reasoning, interacting with environments- it pretty much leapfrogs every other model but coding is the interesting part. vs. Fable 5.1, the gap really isn’t that big

Step 3.7 Flash is now in the Baseten Model Library! This is a 198B-parameter sparse MoE model (11B active per token) with native image and video input, and a 256K context window. It's a strong option for visual reasoning, agentic coding, and long-context tasks.

Kimi K2.5 tech report just dropped! Quick hits: - Joint text–vision training: pretrained with 15T vision-text tokens, zero-vision SFT (text-only) to activate visual reasoning - Agent Swarm + PARL: dynamically orchestrated parallel sub-agents, up to 4.5× lower latency, 78.4% on BrowseComp - MoonViT-3D: a unified image–video encoder with 4× temporal compression, enabling 4× longer videos in the same context - Toggle: token-efficient RL, 25–30% fewer tokens with no accuracy drop Here's our work toward scalable, real-world agentic intelligence. More details in the report 👉

⚡Funding Brief⚡ Annualized revenue reportedly rose from $20 million to $700 million in roughly 12 months. That is the headline metric behind Higgsfield's newly announced $400 million Series B, which values the AI video-and-image company at $5.4 billion — about 4.2 times its reported roughly $1.3 billion Series A and extension valuation in January. The company describes its product as a visual-reasoning platform for creators, brands, agencies and studios: users can generate and develop visual content through AI-driven video and image workflows. Higgsfield was founded by former Snap executive Alex Mashrabov. A few signals worth noting: Higgsfield says business customers now account for a majority of annualized revenue, versus less than one-quarter in January. It also says it provides visual production services to 390 companies in the Fortune 500. These customer metrics are company-reported, and the company has not disclosed customer-spend, retention or contract-value data The company says it has more than 30 million users across 238 countries and territories, with more than 20 million content generations per month Following the May 2026 rollout of Higgsfield Supercomputer, the company says users of its agentic products grew 42-fold over three months DST Global led the round. New investors include Tribe Capital, Growth Equity at Goldman Sachs Alternatives, Smash Capital, Fifth Wall, Valor Capital, Intel Capital, Liberty Global Tech Ventures, Mirae Asset Capital and NTT DOCOMO Ventures; existing investors including Accel and Menlo Ventures also participated The $700 million annualized-revenue figure, user counts, customer coverage and growth data are company-reported operating metrics, not independently audited financial results. For agencies, studios and brands, the commercial question is not whether AI can create content, but whether it can reduce production time and marginal asset costs while meeting quality, brand-control and rights requirements. This financing is a notable signal that investors remain willing to fund AI application-layer companies that report rapid commercial growth. Source: Reuters, PRNewswire, Yahoo Finance, Qz

Making AI "reason about space in words" might be backfiring 🧭 Here's a new approach that lets it imagine unseen viewpoints instead. Title: Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models URL: 🧭 Overview This work proposes Imaginative Perception Tokens (IPT) to strengthen spatial reasoning in vision language models (VLMs). Rather than forcing spatial logic through language, it keeps "what could be perceived under a different arrangement" as an intermediate perceptual representation. ❓ Challenges Solved VLMs struggle with spatial reasoning: inferring unobserved viewpoints, reasoning through occluded paths, and integrating partial observations. Prior work pushed this into textual chain-of-thought, but forcing visual reasoning through language alone hit a ceiling. 💡 Methodology & Proposed Approach ・Uses the unified VLM backbone BAGEL, trained with IPT supervision ・Formulates three tasks: Perspective Taking (PET), Path Tracing (PT), Multiview Counting (MVC) ・Builds a ~20,000-example dataset with ground truth, answers, and metrics The core idea is treating the perception itself ("if I moved here, I'd see this") as an intermediate representation. 📊 Experimental Results ・IPT improved Multiview Counting (MVC) accuracy by 3.4% ・Path Tracing (PT) reached performance competitive with closed-source models ・IPT supervision outperformed textual chain-of-thought training ・Conversely, textual CoT substantially degraded spatial reasoning #SpatialReasoning# #MultimodalLLM#