# Chat-template and tool-declaration determinism as a cache-hit prerequisite > Ollama stores tool properties and tool-call arguments in insertion-ordered maps (`ToolPropertiesMap`, `ToolCallFunctionArguments`). A renderer that loops with `.All()` therefore emits keys in the order the client's JSON had them. Parent: [Mac local LLMs: Prompt cache and persistent KV](https://llms-explorer.com/tree/mac-local-llms-prompt-cache-and-persistent-kv/) · 1 facets · 26 facts · page: https://llms-explorer.com/tree/chat-template-and-tool-declaration-determinism-a/ ## Facts - Ollama stores tool properties and tool-call arguments in insertion-ordered maps (`ToolPropertiesMap`, `ToolCallFunctionArguments`). A renderer that loops with `.All()` therefore emits keys in the order the client's JSON had them. — [source](https://raw.githubusercontent.com/ollama/ollama/main/api/types.go) - The Qwen3.5 renderer writes the tool block first, inside the leading system turn, and serialises each tool with `marshalWithSpaces`. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen35.go) - `marshalWithSpaces` runs `json.Marshal` and then inserts a space after every `:` and `,` that lies outside a string, to mimic Python `json.dumps` spacing. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/json.go) - The Qwen3.5 renderer renders an assistant `` block only for assistant turns after the last real user query, unless the Qwen3.8 variant flag `alwaysRenderAssistantThinkBlock` is set. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen35.go) - The Qwen3-Coder renderer takes the first system message and drops later ones, and adds the default system text "You are Qwen, a helpful AI assistant that can interact with a computer to solve tasks." when tools exist but no system message does. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen3coder.go) - The Qwen3-Coder renderer emits extra schema keys (anything except `type` and `description` on a property, and anything except `type` and `properties` on the parameters object) through `renderAdditionalKeys`. That helper marshals the object, unmarshals it into a `map[string]any`, and ranges over the map. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen3coder.go) - Ollama moved newer models from Go `TEMPLATE` strings to registered renderer and parser pairs; the qwen3-coder renderer carries a TODO to match Python `json.dumps` spacing, and the later Qwen3.5 renderer added `marshalWithSpaces` to do so. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen3coder.go) - llama.cpp issue 26974 (2026-08-12) showed template render time growing quadratically with tool count; PR 27034 merged 2026-08-14 with byte-identical prompts. — [source](https://api.github.com/repos/ggml-org/llama.cpp/pulls/27034) - Go does not define an iteration order for `range` over a map. In `renderAdditionalKeys`, a property that carries two or more extra keys (for example `enum` and `default`) can therefore render those keys in a different order on different requests of the same logical tool set, which breaks the cached prefix. This is inferred from the code and the Go language rule; no issue reporting it was found. — source: `asserted` - The Qwen3.5 renderer has no such map loop for tool schemas, so the hazard is specific to the qwen3-coder renderer path. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen35.go) - Because the tool block precedes the system text and all messages, adding, removing or reordering one tool, or an MCP server reconnecting with its tools in a new order, changes bytes at the start of the prompt and forces a full re-prefill. — source: `asserted` - Insertion-ordered storage makes determinism depend on the client: a client that re-serialises `parameters.properties` with sorted or hash-ordered keys changes the rendered bytes even though the schema is logically equal. — source: `asserted` - Template render time is paid on every request, including full cache hits. At 20 tools of 20 properties the bundled Command-R tool-use template took 0.43 s on llama.cpp before the fix, and 9.2 s at 50 tools of 50 properties. — [source](https://api.github.com/repos/ggml-org/llama.cpp/pulls/27034) - None found. The two engines choose different layers for determinism: Ollama fixes key order in Go structures, llama.cpp relies on the Jinja template and `common_json` insertion order (see llama-cpp-jinja-capability-probes-chat-template.md). — source: `asserted` - Whether `renderAdditionalKeys` ordering has produced an observable cache miss on a real qwen3-coder deployment; no measurement exists. — source: `asserted` - Whether Ollama's prefix reuse compares tokens or text, which decides how much one reordered key costs. — source: `asserted` - Ollama `ToolPropertiesMap` and `ToolCallFunctionArguments` keep insertion order through an ordered map and expose it through `All()`. — [source](https://raw.githubusercontent.com/ollama/ollama/main/api/types.go) - Ollama `marshalWithSpaces` inserts a space after `:` and `,` outside string values after `json.Marshal`. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/json.go) - The Ollama Qwen3.5 renderer renders each tool with `marshalWithSpaces` inside the first system turn, before the system content. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen35.go) - The Ollama Qwen3.5 renderer renders assistant think blocks only for turns after the last user query unless the Qwen3.8 variant is used. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen35.go) - The Ollama Qwen3-Coder renderer ranges over a Go `map[string]any` when it emits extra tool-schema keys. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen3coder.go) - The Ollama Qwen3-Coder renderer keeps only the first system message and injects a default system message when tools exist without one. — [source](https://raw.githubusercontent.com/ollama/ollama/main/model/renderers/qwen3coder.go) - llama.cpp issue 26974 measured chat-template render time of 0.43 s for 20 tools by 20 properties and 9.3 s for 50 by 50 on the Command-R tool-use template. — [source](https://api.github.com/repos/ggml-org/llama.cpp/issues/26974) - llama.cpp PR 27034 (merged 2026-08-14) fixed two quadratic terms in `gather_string_parts` and kept prompts byte-identical. — [source](https://api.github.com/repos/ggml-org/llama.cpp/pulls/27034) - In Ollama's qwen3-coder renderer, two requests with the same tools can render extra schema keys in different orders. — source: `asserted` - A change in tool order or tool set alters the first system turn and invalidates the whole prefix cache on Qwen renderers. — source: `asserted`