Meta details Muse Spark 1.2's multimodal intelligence and visual coding
alexandr_wang · x · 2026-08-21
Meta AI published a research blog detailing the multimodal intelligence of Muse Spark 1.2. The model excels in visual coding, transforming images or videos into functional web pages and games. The blog highlights evaluation results on visual reasoning, chart understanding, and knowledge-intensive tasks. Muse Spark 1.2 shows particularly strong gains when using tools to inspect visual inputs and incorporate them into reasoning, enabling video-heavy workflows common in enterprise use.
More from Multimodal
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- Open-source anatomy-reveal skill turns Claude into an exploded-view motion design studio — msharmas · 2026-10-16
- Flux 3 Klein vs Flux 2 Klein hands-on: the new base model 'barely usable' for T2I — Both-Rub5248 · 2026-10-16
- USTC paper uses speculative sampling to cut latency in interactive video generation — ustc · 2026-10-16
- 800,000 Gaussian splats run in WebGPU; Casado sees near-infinite scaling — generativist · 2026-10-16
- Nano Banana 2.1 image generation now live in Perplexity Computer for all users — ChrisUniverse · 2026-10-16