Multimodal models put different demands on vision encoding, prefill and decoding. Separating vision encoding from the other stages can reduce resource contention and help models respond faster, but only for the right workloads. See how EPD disaggregation works, when it helps and what to consider before using it:

Multimodal reasoning has a latency problem. More video frames leads to more waiting. We built Damage Scout with Gemma 4 on Cerebras, running at over 2,300 toks/s, to show what fast multimodal inference unlocks. Damage Scout samples frames from a rental car walkaround, sends them to Gemma 4, gets back structured findings and box coordinates, then renders an annotated damage report in under 6 seconds. Same task. Same frames. A complete different experience powered by Cerebras ⚡️

A self-improving multimodal model can decide when to switch to fresh problems using only its training logs 🔄 A-RSI: Adaptive Task Refresh for Zero-Data Multimodal Self-Improvement 📌 Overview A-RSI is a method for refreshing the training task pool at the right time when a VLM improves itself without human labels. ⚠️ The problem ・Most work focuses on what to generate next, not when to replace the data ・Labels from majority voting can turn the model's own errors into training targets 🧩 Method ・The base model writes parameterized task generators that programmatically produce an image, a question, and a reference answer ・Training uses GRPO, with correctness ranked above all auxiliary rewards ・When prompts with reward variance drop or accuracy rises, the whole pool is replaced with unused tasks 📊 Results ・Qwen3-VL-8B's 11-benchmark average rose from 57.77 to 61.05 using 800 fresh tasks ・Refreshing every window reached 60.64 but used 3,200 tasks, four times as many ・MiMo-VL-7B also went from 59.08 to 60.71 🔭 Why it matters Timing the refresh without validation scores is practical. Note that each comparison is a single training run. #SelfImprovement# #MultimodalAI#

This new-ish 35B multimodal model went from 3.38 → 20.51 tps on an iPhone. MiniCPM-V 4.7 35B-A3B was updated to decode faster on the iPhone by improving how it moves the model between storage and memory. Constraints 🧠 35B total / ~3B active 📱 iPhone 18 Pro Max — 12GB RAM 👁️ image + multi-image + video understanding Most of those 35B weights don't sit in RAM. The experts can stay in the phone's flash storage and get loaded when they're actually needed. Normally that's slow 🐌 SSD expert loading → 3.38 t/s So OpenBMB built Lookahead Expert Scheduling (LES) It predicts which experts the next layer will probably need and starts loading them while the current layer is still computing. Then they stacked the optimizations: 💾 baseline SSD loading → 3.38 tps 👀 expert lookahead → 6.57 ⚙️ optimized scheduling/loading → 11.34 🗜️ mixed-precision quantization → 17.16 🚀 + DSpark speculative decoding → 20.51 tps That's a 6.07× speedup over the same-phone unoptimized baseline. ⚠️ And these are still OpenBMB's own measurements. Link in ALT

Intern-Decision scales multimodal structured decision-making across 0.8B, 2B, and 4B models. 🤖 🏆 The three models average 79.38, 84.68, and 90.02 across seven decision benchmarks. Intern-Decision-4B surpasses Jev at 88.74 while achieving better probability calibration. ⚡ Reported mean latency is 33.98 ms, 33.28 ms, and 44.16 ms respectively, compared with 109.70 ms for Jev in the same local HF setup. 🧠 One forward pass answers multiple choice, score, and yes/no questions with typed JSON and calibrated probability distributions. 🖼️ Text states and up to eight images enable routing, tool selection, scoring, and multimodal workflow control. 📜 Apache 2.0. Applicable Qwen3.5 upstream notices must be retained.

Enterprise data is inherently multimodal, yet traditional text search leaves valuable insights hidden. Our partnership with @Box changes this by allowing systems to read text and images together. Learn more →

🚛 A Multimodal veteran from the very beginning! Sue Terpilowski of Image Line, has attended every #Multimodal# show since it began. For Sue, the real value is the networking — connecting with people she wouldn’t normally see. Hear more at