Avian.io runs infrastructure for fast, lower-cost inference on open-source large language models, giving developers an OpenAI-compatible API for models like DeepSeek V3.2, Kimi K2.5, GLM-5.1, and MiniMax M2.5 without needing to host the models themselves. Avian Data Inc. was founded in 2022 and operates out of New York.
Rather than a monthly subscription, Avian bills per token consumed with no minimum commitment, positioning the pitch around making frontier open-source models low-cost and fast to deploy compared with running the same models on self-managed infrastructure or paying closed-model API rates. The company frames its core problem as: capable open models exist, but are hard to deploy and higher-priced to run without dedicated infrastructure expertise, which Avian's hosted API is built to remove.
Developers and enterprises building products on open-source LLMs, rather than committing to a single closed-model vendor, are the target users. Pricing runs entirely on consumption, starting around $0.105 per million tokens depending on the model selected, with no subscription fee and no free tier.
