Open models now dominate token volume on Vercel AI Gateway

Melvin Vivas · X post · 2026-09-19 · Open on X

Topics: Industry Trends & Job Market, AI Agents, Tool Use & MCP, LLMOps, Deployment & Monitoring · Level: intermediate

Summary

The creator argues for using open models more to keep costs sustainable, and for improving agent harnesses with Skills to power workflows. He quotes Vercel AI Gateway stats: open models had a record 78.4% share of token volume. Moonshot AI, DeepSeek and Z.ai together outspent OpenAI that day.

Key points

• On Vercel AI Gateway, open models reached 78.4% of token volume vs 21.6% for closed models on a record day.
• By spend, Moonshot AI and DeepSeek ranked #3 and #4; together with Z.ai, their combined spend passed OpenAI (#2).
• Token volume and spend tell different stories: open models are used heavily but cost less.
• The creator recommends moving more workloads to open models to keep costs sustainable.
• Improve agent harnesses with reusable Skills so cheaper models can still run strong workflows.

Resources mentioned

• Vercel · tool · x.com · free · recommended by both Bashiri Smith & Melvin Vivas
  Deploy web apps and frontends.
  Also in: OpenAI DevDay 2026 Recap: Dots, Agents API, Codex Cloud & Marketplace (Melvin Vivas on X · notes), 7 Habits to Become an AI Engineer: Books, Tooling, Research & Shipping (Bashiri Smith on Facebook · notes), Jev model added to the AIBackends API via Vercel AI Gateway (Melvin Vivas on X · notes), Adding Vercel AI Gateway as a provider in AIBackends with Devin (Melvin Vivas on X · notes) and 7 more
• Moonshot AI · tool · moonshot.ai · check price
  Chinese AI lab that makes the open-weight Kimi models.
• DeepSeek · tool · deepseek.com · free
  Open-weight large language models from DeepSeek, known for strong reasoning and coding at low cost.
  Also in: jina-ocr-v1: Turning PDFs, Scans and Tables into Markdown (Melvin Vivas on X · notes), Bolt.new Adds Open Models (GLM, DeepSeek, Kimi) via Bolt Forge (Melvin Vivas on X · notes), OpenCode Go: Open Coding Models (DeepSeek, Qwen, Kimi, GLM) for $10/mo (Melvin Vivas on X · notes), Orchestrator + Subagents in opencode with Open Models (Melvin Vivas on X · notes) and 4 more
• Z.ai (@Zai_org) on X · person · x.com · free
  Official X account of Z.ai, maker of the GLM models and the GLM coding plan.
  Also in: GLM 5.3 Open Weights Release Delayed for Framework Support (Melvin Vivas on X · notes), AI News Roundup: Claude Fable 5, Scientist AI, ZCode, NVIDIA RL (Melvin Vivas on X · notes), Trying ZCode by Z.ai with GLM 5.2 on a Mac (Melvin Vivas on X · notes), Waiting for a Z.ai model to rival Fable (Melvin Vivas on X · notes) and 9 more

Try this

• Use open models more to keep AI costs sustainable.
• Improve your agent harness with Skills to power your workflows.