Vercel ships GLM 5V Turbo
Vercel announced GLM 5V Turbo on its AI Gateway—a vision‑first coding model aimed at converting designs/screenshots into code and supporting LLM‑powered dev tools and observability. The move highlights gateway vendors adding specialized models for developer workflows. (x.com)
Zhipu AI (branded Z.ai) published GLM‑5V‑Turbo on April 1, 2026 as a new native multimodal coding model. (apidog.com) The model is reported at roughly 744 billion parameters in public coverage. (winbuzzer.com) GLM‑5V‑Turbo exposes a very large vision context—reports cite a ~202K token context window geared toward long‑horizon, image+text workflows. (ai-primer.com) Z.ai’s documentation describes GLM‑5V‑Turbo as natively accepting images, video, and text with “native multimodal fusion” intended for planning-plus‑execution agent loops rather than simple image captioning. (docs.z.ai) Vercel added GLM family models to its AI Gateway and multiple outlets list GLM‑5V‑Turbo as available through gateways like Vercel’s, letting teams call the model through a single API endpoint without managing separate upstream provider accounts. (ai-primer.com) Implementation notes from vendors emphasize GLM‑5V‑Turbo’s optimization for agentic workflows (tool invocation, long‑chain execution and OpenClaw scenarios), and early benchmarks published in coverage claim it outperforms Claude Opus 4.5 on agentic browsing tests. (vercel.com)
