[[entry]]
id = "anthropic"
catalog_provider = "anthropic"
display_name = "Anthropic"
endpoint_style = "Anthropic Messages API"
recommended_model = "claude-haiku-4-5-20251001"
recommended_selector = "haiku"
recommended_tool_format = "native"
usage_confidence = "high"
recommended_options = [
'provider = "anthropic"',
'model = "haiku"',
'tool_format = "native"',
'structured_output_mode = "xml_tagged"',
]
notes = [
"Native tools, prompt caching, file upload, and XML-oriented scaffolding are first-class in Harn capability data.",
"Claude 4.7 rows use adaptive thinking; older Claude 4 rows use explicit thinking controls where supported.",
]
caveats = [
"Strict JSON output is modeled as tool-use or XML-tagged output rather than OpenAI-style response_json_schema.",
]
mcp_notes = [
"No provider-specific MCP connector is required; Harn exposes MCP tools through the runtime tool registry.",
]
[[entry]]
id = "openai"
catalog_provider = "openai"
display_name = "OpenAI"
endpoint_style = "OpenAI chat completions / Responses-compatible routes"
recommended_model = "gpt-4o-mini"
recommended_selector = "mid"
recommended_tool_format = "native"
usage_confidence = "high"
recommended_options = [
'provider = "openai"',
'model = "mid"',
'tool_format = "native"',
'structured_output_mode = "native_json"',
]
notes = [
"OpenAI-family routes default to native tool calls and native JSON structured output when the model row supports tools.",
"Reasoning models use developer-role instructions and reasoning-summary transcript projection where the capability row declares it.",
]
caveats = [
"Use explicit reasoning effort only on reasoning rows; non-reasoning chat models should keep thinking disabled.",
]
mcp_notes = [
"Hosted MCP behavior is normalized through Harn tool definitions; provider-side hosted tools remain a separate provider feature.",
]
[[entry]]
id = "gemini"
catalog_provider = "gemini"
display_name = "Gemini API"
endpoint_style = "Gemini generateContent"
recommended_model = "gemini-2.5-flash"
recommended_selector = "gemini:gemini-2.5-flash"
recommended_tool_format = "native"
usage_confidence = "medium"
recommended_options = [
'provider = "gemini"',
'model = "gemini-2.5-flash"',
'tool_format = "native"',
'structured_output_mode = "native_json"',
]
notes = [
"Harn lowers native tools to Gemini function declarations and maps function responses back into the transcript.",
"Gemini response usage maps cached-content token counts when the provider reports them.",
]
caveats = [
"Harn does not create Gemini context-cache resources yet; cache accounting is therefore observational.",
]
mcp_notes = [
"MCP tools are regular Harn runtime tools before they become Gemini function declarations.",
]
[[entry]]
id = "mistral"
catalog_provider = "openrouter"
display_name = "Mistral via OpenRouter"
endpoint_style = "OpenAI-compatible chat completions through OpenRouter"
recommended_model = "mistralai/mistral-small-2603"
recommended_selector = "openrouter:mistralai/mistral-small-2603"
recommended_tool_format = "native"
usage_confidence = "medium"
recommended_options = [
'provider = "openrouter"',
'model = "mistralai/mistral-small-2603"',
'tool_format = "native"',
]
notes = [
"Harn catalogs hosted Mistral routes through OpenRouter today, so endpoint and auth behavior are OpenAI-compatible.",
"Use this row for Mistral-family recommendation surfaces until a direct Mistral provider is cataloged.",
]
caveats = [
"Provider-native behavior depends on the OpenRouter model route; run the coding-agent benchmark before promoting it to a default for critical harnesses.",
]
mcp_notes = [
"MCP tools are rendered as OpenAI-compatible tool definitions on this route.",
]
[[entry]]
id = "ollama"
catalog_provider = "ollama"
display_name = "Ollama"
endpoint_style = "Ollama native chat API"
recommended_model = "devstral-small-2:24b"
recommended_selector = "devstral-small-2"
recommended_tool_format = "text"
usage_confidence = "high"
recommended_options = [
'provider = "ollama"',
'model = "devstral-small-2"',
'tool_format = "text"',
'thinking = "off"',
]
notes = [
"Local Ollama model quality varies by template and quantization; Harn defaults known fragile routes to the text-tool contract.",
"Use `harn provider-tool-probe` receipts to promote aliases from unknown to native/text/disabled on a machine.",
]
caveats = [
"Some Ollama native tool parsers reject otherwise valid text-mode model output; the capability table records those routes as text-only.",
]
local_setup_notes = [
"Run `harn models install devstral-small-2` or `harn models install qwen3.6-coding`, then verify with `harn provider-ready ollama --model <model>`.",
]
mcp_notes = [
"MCP tools are local Harn tools; prefer the text tool contract unless a probe proves native calls work for the installed model.",
]
[[entry]]
id = "local"
catalog_provider = "local"
display_name = "OpenAI-compatible local server"
endpoint_style = "OpenAI-compatible chat completions"
recommended_model = "gemma-4-26b-a4b-it"
recommended_selector = "local-gemma4"
recommended_tool_format = "text"
usage_confidence = "low"
recommended_options = [
'provider = "local"',
'model = "local-gemma4"',
'tool_format = "text"',
]
notes = [
"Use this generic provider when a local server speaks OpenAI chat completions but does not need a provider-specific quirk profile.",
]
caveats = [
"Prefer `llamacpp` or `mlx` when those runtimes are known, because their capability rows can encode template-specific behavior.",
]
local_setup_notes = [
"Set `LOCAL_LLM_BASE_URL` and either `LOCAL_LLM_MODEL` or an explicit Harn model selector, then run `harn provider-ready local`.",
]
mcp_notes = [
"MCP tools are exposed as OpenAI-compatible tool definitions unless the route is configured to prefer Harn text tools.",
]
[[entry]]
id = "llamacpp"
catalog_provider = "llamacpp"
display_name = "llama.cpp server"
endpoint_style = "OpenAI-compatible llama-server"
recommended_model = "qwen3.6-35b-a3b-ud-q4-k-xl"
recommended_selector = "llamacpp-qwen3.6-q4"
recommended_tool_format = "text"
usage_confidence = "medium"
recommended_options = [
'provider = "llamacpp"',
'model = "llamacpp-qwen3.6-q4"',
'tool_format = "text"',
'thinking = "off"',
]
notes = [
"llama.cpp gets its own provider so Harn can model Qwen chat-template and thinking behavior separately from generic local OpenAI-compatible servers.",
]
caveats = [
"Run both provider readiness and tool probes after changing GGUF, context, KV-cache, or chat-template settings.",
]
local_setup_notes = [
"Run `harn models install local-qwen3.6-gguf` for the recommended download and launch commands.",
]
[[entry]]
id = "mlx"
catalog_provider = "mlx"
display_name = "MLX OpenAI-compatible server"
endpoint_style = "OpenAI-compatible MLX server"
recommended_model = "unsloth/Qwen3.6-27B-UD-MLX-4bit"
recommended_selector = "mlx-qwen3.6-27b"
recommended_tool_format = "native"
usage_confidence = "medium"
recommended_options = [
'provider = "mlx"',
'model = "mlx-qwen3.6-27b"',
'tool_format = "native"',
]
notes = [
"MLX routes use native tools only after the served identity and tool probe match the cataloged model.",
]
caveats = [
"`mlx-vlm` server flags vary by release; launch first, then verify with `harn provider-ready mlx`.",
]
local_setup_notes = [
"Run `harn models install local-qwen3.6-27b` for the venv, download, launch, and verification commands.",
]