models-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_providersA | List every provider in the models.dev catalog (e.g. anthropic, openai, google) with its model count. Use this to see what's available before narrowing with find_models. |
| find_modelsA | Search and filter models across all providers by name, provider, capability, context window, or price. Returns a compact summary per model; use get_model for full details on one. |
| get_modelA | Fetch full metadata for one model. Accepts 'provider/model' (preferred, e.g. 'anthropic/claude-sonnet-4-5') or a bare model id if unambiguous across providers. |
| compare_modelsA | Diff 2-6 models side by side on pricing, context window, and capabilities. Accepts 'provider/model' ids, or bare model ids when unambiguous across providers. Ids resolving to the same model are compared once. |
| refresh_catalogA | Force a fresh fetch of the models.dev catalog, bypassing the in-memory cache. Use if you suspect the data is stale. If the refetch fails, the last good catalog keeps being served and servedStale is reported as true. |
| estimate_costA | Estimate the USD cost of a request to one model from its published per-million-token rates. Use get_model first if you need to pick an id. |
| top_modelsA | Rank models by one criterion: cheapest input/output price per million tokens, largest context window, most context per input dollar, or newest release. Supports the same filters as find_models. |
| get_providerA | Fetch metadata for one provider: display name, AI SDK package name, API base URL, docs link, and a compact list of its models. Accepts the provider id ('anthropic') or display name ('Anthropic'). Use get_model for full per-model detail. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 8 tools
Each tool targets a distinct operation: listing providers, searching models, fetching details, comparing, refreshing, estimating cost, ranking, and provider metadata. Even where two tools support similar filters, their purposes are clearly differentiated by descriptions.
Tool names mostly follow a clear verb_noun pattern such as list_providers, find_models, get_model, and estimate_cost. The exception is top_models, which reads more like a noun phrase than an imperative verb action, a minor deviation from the otherwise consistent convention.
Eight tools is a well-scoped size for a model catalog server. Each tool covers a meaningful workflow without redundancy or bloat, from discovery and search to comparison and cost estimation.
The surface covers the full read-only lifecycle of interacting with the models.dev catalog: list, search, detail, compare, rank, and cost estimation, plus provider-level metadata and cache refresh. No obvious dead ends or missing core operations are present.