Truly open weights, and the licenses that only look open
"Open weight" covers everything from Apache-2.0 — self-host it, ship it in a product, no strings — to licenses that forbid commercial use or bolt on acceptable-use clauses. Across 107 models here, 51% are permissively licensed (Apache / MIT), 29% are open but conditional, and 20% are proprietary — weights closed, API only. A license belongs to the model, identical whichever provider serves it, so each figure below is one host per model (the cheapest).
Cheapest genuinely open model: mistral-nemo ships Apache 2.0 weights — commercial use and self-hosting, no clause — at $0.049/1M blended (DeepInfra).
Permissive means the license itself won't stop you shipping it. It says nothing about running cost — see output vs inputand best value per dollar for that.
How open is the field?
Share of models in each license tier.
Licenses in the index
| License | Tier | Models | Reference |
|---|---|---|---|
| Apache 2.0 | permissive | 33 | terms → |
| MIT | permissive | 22 | terms → |
| Proprietary | proprietary | 21 | terms → |
| Llama Community License | conditional | 6 | terms → |
| MiniMax Model License | conditional | 6 | terms → |
| Modified MIT | conditional | 5 | terms → |
| Mistral Research License | conditional | 4 | terms → |
| NVIDIA Open Model License | conditional | 4 | terms → |
| Gemma Terms of Use | conditional | 3 | terms → |
| CC-BY-NC 4.0 | conditional | 2 | terms → |
| Mistral Non-Production License | conditional | 1 | terms → |
Cheapest truly-open models
Permissively licensed (Apache / MIT) — self-host, run commercially, no clause — ranked by price at their cheapest host.
| Model | License | Cheapest host | Blended /1M |
|---|---|---|---|
| glm-5.2 | MIT | OpenRouter | $0 |
| mistral-nemo | Apache 2.0 | DeepInfra | $0.049 |
| gpt-oss-20b | Apache 2.0 | AkashML | $0.12 |
| deepseek-v4-flash | MIT | Baidu | $0.15 |
| gpt-oss-120b | Apache 2.0 | AkashML | $0.2 |
| qwen3.5-9b | Apache 2.0 | SiliconFlow | $0.25 |
| granite-4.2-8b | Apache 2.0 | CoreWeave | $0.25 |
| mistral-small-3.2-24b | Apache 2.0 | DeepInfra | $0.275 |
| glm-5.3-flash | MIT | Relace | $0.309 |
| qwen3-coder-30b-a3b | Apache 2.0 | Novita AI | $0.34 |
| qwen3-14b | Apache 2.0 | DeepInfra | $0.36 |
| qwen3-32b | Apache 2.0 | DeepInfra | $0.36 |
| qwen3-30b-a3b | Apache 2.0 | SiliconFlow | $0.39 |
| voxtral-small-24b | Apache 2.0 | Mistral AI | $0.4 |
| qwen3-235b-a22b | Apache 2.0 | GMICloud | $0.438 |
| glm-4.7-flash | MIT | Venice | $0.46 |
| deepseek-v3.2 | MIT | GMICloud | $0.518 |
| qwen3-vl-8b | Apache 2.0 | Alibaba | $0.572 |
| qwen3-vl-30b-a3b | Apache 2.0 | Alibaba | $0.65 |
| deepseek-v3.2-exp | MIT | SiliconFlow | $0.68 |
Prices per 1M tokens (USD), each model's cheapest published rate. License is the model family's reference license — see each terms link. Verified 2026-09-05.
FAQ
What does 'open weight' actually mean for an LLM?
It ranges widely. Across 107 models here, 51% are permissively licensed (Apache/MIT — self-host and ship commercially, no strings), 29% are open but conditional (non-commercial or acceptable-use clauses), and 20% are proprietary — weights closed, API only.
What's the cheapest truly-open model I can self-host and ship?
mistral-nemo ships Apache 2.0 weights — commercial use and self-hosting with no clause — at $0.049 per 1M tokens blended via DeepInfra, the cheapest genuinely open option in the index.
Is a permissive license enough to use a model in production?
The license only settles whether you're allowed to ship it — permissive means it won't stop you. It says nothing about running cost or quality; check the value ranking and the model's per-token price separately.