--- title: Does OpenCode Load All MCP Tools? OpenCode 1 vs 2, Measured description: OpenCode 1 sent 63,245 tokens of MCP tools with every request; OpenCode 2 showed 16 of 99 tools in 881 tokens. Measured on a 99-tool server, plus Grok Build. canonical: https://spiritroom.app/blog/opencode-mcp-tool-loading/ published: 2026-10-21 updated: 2026-10-21 --- > Markdown version of https://spiritroom.app/blog/opencode-mcp-tool-loading/ for agents. Index: /llms.txt [Blog](/blog/) / [How agents load MCP tools](/guides/how-agents-load-mcp-tools/) # OpenCode 1 vs OpenCode 2: from a 63k-token tool preload to a 16-tool catalog OpenCode is the open-source agent app many teams pick when they want their own model. Its two major versions load MCP tools in opposite ways. If your users run it, that decides most of what your server costs them. Leonid Bugaev · Published 21 Oct 2026 **TL;DR** - **OpenCode 1.18:** all 99 tools, 63,245 tokens, sent with every model request. About 230k tokens of definitions per question. - **Cost:** 341,615 input tokens per question on average, about 1.6× Codex or Claude Code. Answers were still right: 39/40 (87–100%). - **OpenCode 2.0 (opt-in):** a 16-of-99 catalog, 881 tokens, and a `search()` inside its code tool for the rest. 2 of 2 live answers correct. - **Both cut large results** at 2,000 lines or 51,200 bytes, keeping the start and saving the rest to a file. - **Grok Build** preloads nothing: it lists server names and counts and searches tools with BM25. [OpenCode 1: everything, every time](#oc1)[OpenCode 2: a partial catalog plus search](#oc2)[Result cut-offs in both](#cap)[Sidebar: Grok Build](#grok)[What it means for your server](#means)[How we measured](#how-we-measured)[FAQ](#faq) ## OpenCode 1: everything, every time We captured the first request OpenCode 1.18.34 sends to its model with a 99-tool production MCP server attached, both live and with a recorded copy of the tool list. Both gave the same result. OpenCode 1: MCP definitions per request *63,245 tokens (99 tools)* OpenCode 2: catalog per request *881 tokens (16 tools)* OpenCode 1 sends the definitions again with every model request, not just the first. In 40 live runs it averaged 3.7 requests per question, so about **230k tokens of tool definitions per question** (3.7 × 63,245, an estimate). - Average input per question: **341,615 tokens** by OpenCode's own usage figures, about 1.6× Codex or Claude Code on the same jobs. - Largest single request: 131,714 tokens, 13% of its model's 1M limit. Big, but it fit. - Answers: **39/40** correct (87–100%). Preloading cost tokens, not accuracy. **See what your MCP server costs in OpenCode.** We run your users' jobs in OpenCode, Claude Code, Codex and Grok Build, and show what each loads and pays for. [Test your MCP →](/#contact) ## OpenCode 2: a partial catalog plus search OpenCode 2.0.31 hides MCP tools behind one code-running tool. In 2 live read-only runs in the interactive app, the model saw: ``` - server (99 tools, 16 shown) The catalog is partial. ... search() ... ``` - 16 of the 99 tools with one-line descriptions, 881 tokens; the first request was about 5,270 tokens. - The model called `search()` from code to find the others. Both answers were correct (2/2, 34–100%). - Headless `run` mode never showed the MCP tools to the model in our tests; we didn't confirm why. With a catalog, what the model sees first is a short description line, so the first words of each description decide a lot. See [put the meaning in the first line](/blog/mcp-tool-descriptions-first-line/). ## Result cut-offs in both Both versions keep the first **2,000 lines or 51,200 bytes** of a result and save the full output to a file. In OpenCode 1, **29 of 33** large list results were cut. In OpenCode 2 one job hit the cap twice (14,098 and 16,343 tokens kept), then the model grepped and read the saved file over 4 steps and answered correctly. More: [output limits by agent](/blog/mcp-tool-output-limits-by-agent/). ## Sidebar: Grok Build - First request: 28 built-in tools, 19,674 tokens. No MCP tools on the wire. - A reminder after the user's message lists connected servers and tool counts, and says to call `search_tool` before `use_tool`. - `search_tool` runs BM25 on the client, default limit 5. - A project-scoped server didn't start in an untrusted folder, and headless mode gave no warning. ## What it means for your server - **For OpenCode 1 users,** every tool you add is resent on every request. Trim definitions and split rarely used tools. See [how many MCP tools is too many](/blog/how-many-mcp-tools-is-too-many/). - **For OpenCode 2 and Grok Build,** names and first lines decide discovery. See [name tools the way users talk](/blog/mcp-tool-naming-for-agent-search/). - **For both OpenCode versions,** keep results under 51,200 bytes or page them. Context in general: [MCP context window usage](/blog/mcp-context-window-usage/). Back to [how agents load MCP tools](/guides/how-agents-load-mcp-tools/). **Go deeper:** [Claude Code tool search, seen from the wire](/blog/claude-code-tool-search-explained/), [How Codex finds MCP tools](/blog/how-codex-finds-mcp-tools/), [MCP tool-list token calculator](/tools/mcp-token-cost/). ## How we measured - **OpenCode 1.18.34:** first-request captures (live and recorded list), 40 live read-only runs on 16 Oct 2026 with its owner's default model; tokens by o200k\_base for definitions, OpenCode's own usage for totals. - **OpenCode 2.0.31:** 2 live runs driven through the interactive app, 16 Oct 2026. Very small sample. - **Caps:** read from the shipped binaries; cut-offs counted in transcripts. - **Grok Build 1.0.60:** one captured task against a 99-tool stub. - **Limits:** versions change fast; one server; ranges are 95% Wilson intervals. ## FAQ ### Does OpenCode load all MCP tools into context? OpenCode 1.x does: in our capture, all 99 tools (63,245 tokens) went into every model request. OpenCode 2.0 doesn't by default: it showed a catalog of 16 of the 99 tools and let the model search for the rest. ### Which OpenCode version do users have? In our tests OpenCode's standard installer still installed version 1; version 2 was a separate opt-in release. ### How does OpenCode 2 choose which tools to show? Its catalog said it was partial. The 16 shown appeared to be those with the cheapest schemas; the model used search() from code to find others. ### What does OpenCode do with large tool results? Both versions keep the first 2,000 lines or 51,200 bytes and save the full output to a file the model can read. In OpenCode 1, 29 of 33 large list results were cut. ### How does Grok Build load MCP tools? It sends none up front. A reminder lists connected servers and tool counts; the model calls search\_tool (client-side BM25, default limit 5) and then use\_tool. **See what your MCP server costs in OpenCode.** We run your users' jobs in OpenCode, Claude Code, Codex and Grok Build, and show what each loads and pays for. [Test your MCP →](/#contact) ## Keep reading - Up[How agent apps load MCP tools: what the model sees (guide)](/guides/how-agents-load-mcp-tools/) - Related[How many MCP tools is too many? Real token costs across five agents (2026)](/blog/how-many-mcp-tools-is-too-many/) - Related[Why your MCP server fills the context window, and which agents actually pay for it](/blog/mcp-context-window-usage/) - Related[MCP output limits by agent: Claude Code, Codex, OpenCode and Grok Build compared](/blog/mcp-tool-output-limits-by-agent/)