Runs onThe places the vendor's own pages say the product runs: your machine, a container environment you run, or a cloud service the vendor runs. It is not a limit.
Models runWhere the vendor's own pages say the product's models run, not a limit: your machine, a container environment you run, the model company's cloud (servers of the company that makes or serves the model), or a cloud provider's model service. For a cloud, read the data and compliance lines of the company running it, linked by its name.
Several providers in one sessionOne session can use models from different providers, for example one agent on one maker's model and another on a second maker's. Switching the single model for the whole session does not count on its own. A tool that is itself a gateway or library routing to the providers its page names counts.
Several endpoints built inThe tool itself can hold more than one provider or address, so nothing has to sit in front of it to reach different providers.
Local and cloud togetherA local model (on your machine or in a container environment you run) and a cloud model (in its model company's cloud or a cloud provider's model service) can both be used in the same session or setup.
Use a local model: How to set it up(opencode.ai page, read 2026-10-19) Settings the guide names: baseURL, options, provider.
Use a local and a cloud model together: How to set it up(opencode.ai page, read 2026-10-19) Settings the guide names: provider, model, opencode.json.
Hold several endpoints in the product: How to set it up(opencode.ai page, read 2026-10-19) Settings the guide names: baseURL, options, provider, opencode.json.
LiteLLM: runs OpenCode on any model through its Python SDK (github.com)
Compared with similar tools
Same role (harness / agent client) at the same layer (Coding agents and assistants). Up to 4 others, in alphabetical order; nothing is ranked.
Product
Several providersOne session can use models from different providers, for example one agent on one maker's model and another on a second maker's. Switching the single model for the whole session does not count on its own. A tool that is itself a gateway or library routing to the providers its page names counts.
Several endpointsThe tool itself can hold more than one provider or address, so nothing has to sit in front of it to reach different providers.
Model on your machineThe tool can use a model running on your own computer, for example through Ollama, LM Studio, oMLX or llama.cpp.
Local and cloudA local model (on your machine or in a container environment you run) and a cloud model (in its model company's cloud or a cloud provider's model service) can both be used in the same session or setup.
Words in the cells: YesA page we read says so., PartlyA page says part of it, for example separate models per agent without saying the providers can differ., Via a proxyOnly with a proxy, or a local server that imitates a cloud provider's interface, in front of the tool., NoA page we read says the product does not support it., ClaimedA vendor page says so on a marketing or overview page, and we found no guide showing how to set it up., Not foundWe read pages for it and found nothing there. It says nothing about whether the product has the feature., n/aThe question does not apply to the product, for example because it is a model server and not a tool that connects to models., BlankNothing is published for it yet.
Data and compliance
Compliance: Set by whoever runs itThe answer belongs to whoever runs the software, because open-source software has no maker running it for you., for their own deployment.
Data: Set by whoever runs itThe answer belongs to whoever runs the software, because open-source software has no maker running it for you., with telemetry not checked yetWe have not read a page of the right kind for this question yet..
Model and where it runs: Run by its operator.
Details: What the project's own pages state, not a compliance determination.
What the maker's own pages said on the dates shown, not a compliance determination.
GitHub page: opencode(the product's own site, from our list of products; not read for these lines)
The readings behind this page
Our summary of the vendor's pages as of the dates shown on each line (2026-10-19 to 2026-10-24); confirm on the vendor's page.
Local models
OpenCode's page says it supports 75+ LLM providers and running local models. (opencode.ai page, read 2026-10-19)
Cloud models
OpenCode's models page says most popular providers come preloaded by default, and a provider whose credentials were added with the connect command is available at startup. (opencode.ai page, read 2026-10-19)
Local and cloud models in one session
OpenCode lets the user define specialized agents for specific tasks, and the page's example gives such an agent its own model entry. (opencode.ai page, read 2026-10-19)
What you set up
In OpenCode's providers page, the examples for local model servers (Atomic Chat, LM Studio and Ollama among them) say the options.baseURL setting is the endpoint of that local server. (opencode.ai page, read 2026-10-19)
OpenCode's models page says a default model is set with the model key in the OpenCode config, written as provider ID and model ID; its example sets a model under the lmstudio provider. (opencode.ai page, read 2026-10-19)
Models from different providers in one session
In OpenCode, an agent's model config overrides the model for that agent, which the page says is useful for matching different tasks with models suited to them. (opencode.ai page, read 2026-10-21)
Several endpoints, built in (no proxy)
OpenCode lets you change the base URL of any provider through its baseURL option. (opencode.ai page, read 2026-10-21)
What a local model must meet
OpenCode's models page says models discovered from vLLM advertise no tool support, because discovery does not report the server flags that switch tool calling on. (opencode.ai page, read 2026-10-23)
OpenCode's providers page says models found by vLLM discovery start with tools switched off, because discovery cannot tell whether a model supports tools. (opencode.ai page, read 2026-10-23)
OpenCode's models page says that for a model not already in its catalog, it assumes a context limit of 200,000 tokens and an output limit of 32,000 tokens. (opencode.ai page, read 2026-10-23)
OpenCode's models page says that for vLLM it takes the context limit from the max_model_len the server reports. (opencode.ai page, read 2026-10-23)
Plan, policy and billing
OpenCode's providers page describes OpenCode Go as a subscription you can choose to take, covering coding models that the OpenCode team has tested. (opencode.ai page, read 2026-10-23)
OpenCode's providers page says a ChatGPT Pro or Plus plan can replace an OpenAI API key. (opencode.ai page, read 2026-10-23)
OpenAI's Sign in with ChatGPT page names OpenCode among four open-source integrations, the others being OpenClaw, Pi and T3. (learn.chatgpt.com page, read 2026-10-23)
OpenAI's help article on using a ChatGPT plan in other apps says that, for the open-source tools it supports, availability extends to all ChatGPT users. (help.openai.com page, read 2026-10-23)
On OpenCode's providers page, the Cloudflare AI Gateway entry says that with Cloudflare's Unified Billing a reader does not need a separate API key for each provider. (opencode.ai page, read 2026-10-19)
A separate model for planning and for doing
The model config lets you override the agent's model. (opencode.ai page, read 2026-10-21)
OpenCode's agents page gives, as an example of assigning different models to different agents, a quicker model for planning and a stronger one for implementation. (opencode.ai page, read 2026-10-21)
A separate model is configured by the small_model option for lightweight tasks such as title generation. (opencode.ai page, read 2026-10-19)
OpenCode provides two primary agents, Build and Plan, as built-in features. (opencode.ai page, read 2026-10-21)
About the product
OpenCode's README says OpenCode also comes as a desktop application, which it marks as a beta. (github.com page, read 2026-10-24)
Our reading of vendors' pages, as of the dates shown. Confirm on the vendor's page before relying on any line that cites it. To report an error write to report@crinaro.ai