AllocateMIT licensed language models, ranked by price
The catalog lists 9 MIT licensed language models, from DeepSeek R1 Distill Qwen 1.5B at $0.18 per million input tokens to DeepSeek’s reasoning line. MIT is the most permissive common license: use, modify, fine-tune, and redistribute, commercially, with attribution as the only obligation.
Provider list prices from the Allocate catalog, checked 2026-07-08. Billed price on Allocate is list plus the 7% transaction fee.
MIT in practice
MIT asks one thing: keep the license notice. There are no usage thresholds, no field-of-use limits, and no conditions on derivative weights, which makes it the simplest possible answer to a procurement review.
The list ranks by combined list price. The GLM 4.x line and DeepSeek’s open releases dominate it; both take fine-tuning well and serve at open-market prices.
Common questions
What does MIT allow?
Commercial use, modification, distribution, and private use, with attribution. Fine-tuned weights derived from an MIT base carry no conditions from the license.
What is the cheapest MIT model?
DeepSeek R1 Distill Qwen 1.5B at $0.18 per million input tokens at list.
Related
Every model here sits behind one key on Allocate: route by name, meter per route, and swap the model in one click.