Ollama

Ollama is the command-line tool and desktop app that lets you download and run open models such as Qwen, Gemma, DeepSeek and gpt-oss on your own machine, with a local REST API that coding tools and chat apps plug into. It is MIT-licensed, and running models on your own hardware is free without limits.

What the free plan gives you

The Free plan on ollama.com lists running models locally, access to starter models, starter usage credits for Ollama's cloud-hosted models and no service fees. The pricing page states that running models on your own hardware is always unlimited, so the only cost of local inference is your GPU or CPU time. Nothing is metered on the local app: unlimited models, prompts, tokens and API calls, and your data never leaves the machine.

Where the free plan stops

The paid tiers only matter if you want Ollama to run large models in its cloud for you. The Pro plan is $20 a month ($200 a year) and unlocks all cloud models with a higher usage allowance; Max is $100 a month, and Team is $500 a month in early access. The Free plan's starter credits cover a taste of cloud models, after which you either add credits or stick to local. There is no local feature that is gated behind payment.

Who it suits

Developers who want a local LLM endpoint for coding assistants, privacy-sensitive automation or offline use should install it and never think about the pricing page. If you lack a capable GPU and want to run 100B-plus parameter models, Pro's cloud access is the paid path, though at that point you are also comparing against hosted API providers.