Ling 3.0 Tiny is now available on AI Gateway
Alibaba's Ling 3.0 Tiny lands on Vercel's AI Gateway—free through August 13, built for agentic workloads.
Why it matters
A capable small open-weight model (1.3B active params, 256K context, native function calling) joins a unified API platform, lowering the bar for practitioners to test agent-oriented inference without vendor lock-in or platform markup.
The key facts
12 to knowLing 3.0 Tiny: 7.9B total parameters, ~1.3B active per token (MoE)
256K context window, up to 32K output tokens
Native function calling and prompt caching
Free access on Vercel AI Gateway through August 13, 2026
Model designed for responsive agents, instruction following, multi-turn conversation
AI Gateway: no platform inference fee, no markup on provider pricing, includes cost tracking, routing, failover, retries
After August 13, model available as inclusionai/ling-3-0-tiny (paid tier)
Ling 3.0 Tiny: 7.9B parameters, 1.3B active per token (MoE)
Built for responsive agents, instruction following, multi-turn conversation
AI Gateway: unified API with usage tracking, cost management, retries, failover, routing rules
Zero platform fee on inference, including BYOK requests
Model provider: ANT Group
Go to the source
Vercel Blogvercel.com
Publisher excerpt: from ANT Group is now on AI Gateway, free to use till August 13. Ling 3.0 Tiny takes the free slot from .Ling 3.0 TinyLing 3.0 Flash Ling 3.0 Tiny is a MOE model with 7.9B total parameters and about 1.3B active per token, a 256K token context window, and up to 32K output tokens. The model is built…