The article reviews DeepSeek‑V4‑Flash‑Vision‑Exp, exposing a misidentification bug, detailing its visual‑primitive approach, impressive spatial‑reasoning benchmarks, and a highly compressed KV‑cache architecture that balances performance with efficiency.
The TianMou team at Tsinghua University expands its Nature‑cover breakthrough from a novel brain‑inspired complementary vision chip to a full self‑supervised algorithmic ecosystem that learns visual primitives without ground‑truth, delivers high‑dynamic‑range imaging, and powers downstream tasks such as depth estimation and video segmentation in extreme open‑world conditions.
The article details how Tsinghua University's brain‑inspired complementary vision paradigm, embodied in the TianMouChip and its self‑supervised IGFNet framework, tackles visual degradation in open‑world settings, delivering high‑quality perception with low hardware overhead and enabling robust downstream tasks such as depth estimation and video segmentation.
DeepSeek’s new paper "Thinking with Visual Primitives" tackles the reference gap in multimodal models by introducing points and boxes as reasoning units, achieving up to 8× token efficiency and leading benchmark scores in counting, spatial reasoning, and maze navigation compared with GPT‑5.4, Claude‑Sonnet‑4.6 and Gemini‑3‑Flash.
The AI roundup covers OpenAI's Codex upgrade with Workspace Agents and 40% token efficiency, xAI's Grok 4.3 API offering 128K context and 60% lower pricing, Ant Group's open‑source Ling 2.6‑1T model, DeepSeek's multimodal Visual Primitives framework and its sudden removal, plus the ongoing GPT‑Plus account bans and their mitigation.
DeepSeek’s multimodal model, built on the V4‑Flash architecture and a visual‑primitive reasoning approach, compresses a full‑resolution image by 7,056 times, achieves comparable or superior performance to GPT‑5.4 and Claude‑Sonnet‑4.6 on counting and spatial‑reasoning benchmarks, and does so with dramatically lower compute.
DeepSeek releases an open‑source multimodal LLM that introduces a visual‑primitive framework—elevating bounding boxes and points to token level—to close the reference gap, achieve extreme KV‑cache compression, and outperform GPT‑5.4, Claude‑Sonnet‑4.6 and Gemini‑3‑Flash on counting, spatial reasoning, maze navigation and path‑tracing benchmarks.
DeepSeek has released a multimodal model built on a visual‑primitive reasoning paradigm that treats coordinates and bounding boxes as reasoning units, dramatically compresses visual tokens, and achieves state‑of‑the‑art performance on counting, spatial, and topological tasks, while exposing current limits of multimodal inference.
