The Aggregate — LLM benchmark aggregate is an MCP server with 8 tools an agent can call, among them about_the_aggregate, compare_models, get_benchmark. It answers at https://theaggregate.ai/mcp over streamable HTTP, and it listed its tools to an anonymous client, so no sign-in was needed for that. Our probes reached it 10 of 10 times since 4 September 2026; the last check, on 17 October 2026, succeeded.
Fused LLM rankings: one IRT/Elo scale across ~5,000 public benchmark leaderboards, updated daily.
Description as the publisher wrote it.
| Tool | What it does |
|---|---|
about_the_aggregate | What this data is: how the IRT fusion works, current coverage counts, update cadence, and how to cite it. |
compare_models | Head-to-head between 2-4 models: aggregate ranks, Elo gap with a significance note based on the standard errors, and notable benchmarks they share. Inputs: models |
get_benchmark | One benchmark in depth: what it measures, the original source leaderboard URL, IRT stats (difficulty, noise, model coverage), skill weights, and the current top models on it. Inputs: benchmark, top |
get_leaderboard | Top of the cross-benchmark aggregate ranking: every model placed on one Elo scale by an IRT model fit over ~6,300 public benchmark leaderboards. One row per model by default, fused across reasoning-effort settings. Supports paging via limit Inputs: limit, offset, include_variants |
get_model | One model in depth: aggregate rank, Elo with standard error, provider, what it is, cost per task where known, and its most notable benchmark results (with percentiles). Inputs: model |
get_prediction_duel | Guesswork — the public prediction duel: every day frontier LLMs and The Aggregate's own IRT model predict newly scraped benchmark scores before seeing them, and the errors are scored. Returns the current monthly standings, wins and losses i |
search_benchmarks | Find benchmarks in the aggregate by (partial) name. Returns model coverage, difficulty on the Elo scale, and the benchmark page URL. Inputs: query, limit |
search_models | Find ranked models by (partial) name or provider. Returns rank, Elo and the model page URL. One row per model by default, fused across reasoning-effort settings. Inputs: query, limit, include_variants |
Every block below points at https://theaggregate.ai/mcp. Pick your client and copy.
claude mcp add --transport http the-aggregate-llm-benchmark-aggr https://theaggregate.ai/mcp
{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://theaggregate.ai/mcp"]
}
}
}{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"url": "https://theaggregate.ai/mcp"
}
}
}{
"servers": {
"the-aggregate-llm-benchmark-aggr": {
"type": "http",
"url": "https://theaggregate.ai/mcp"
}
}
}{
"context_servers": {
"the-aggregate-llm-benchmark-aggr": {
"url": "https://theaggregate.ai/mcp"
}
}
}{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"serverUrl": "https://theaggregate.ai/mcp"
}
}
}{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"url": "https://theaggregate.ai/mcp"
}
}
}name: The Aggregate — LLM benchmark aggregate
version: 0.0.1
schema: v1
mcpServers:
- name: The Aggregate — LLM benchmark aggregate
type: streamable-http
url: https://theaggregate.ai/mcp{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"httpUrl": "https://theaggregate.ai/mcp"
}
}
}{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"type": "http",
"url": "https://theaggregate.ai/mcp"
}
}
}{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"url": "https://theaggregate.ai/mcp"
}
}
}goose://extension?url=https%3A%2F%2Ftheaggregate.ai%2Fmcp&type=streamable_http&id=the-aggregate-llm-benchmark-aggr&name=The%20Aggregate%20%E2%80%94%20LLM%20benchmark%20aggregate
It answered our last check on 17 October 2026 in 279 ms. Over 10 checks since 4 September 2026 it answered 10. Status here comes from Neuronto's own probes, not from the publisher.
Yes. It is published by theaggregate.ai and served from theaggregate.ai, the vendor's own domain: https://theaggregate.ai/mcp.
8, as returned by the server's own tools/list on 8 September 2026: about_the_aggregate, compare_models, get_benchmark, get_leaderboard, get_model, get_prediction_duel, search_benchmarks, search_models.
It answered our last check on 17 October 2026 in 279 ms. Over 10 checks since 4 September 2026 it answered 10.
Put the server under mcpServers in ~/.cursor/mcp.json with the URL https://theaggregate.ai/mcp. The exact JSON is above, with the same setup for Claude Code, Claude Desktop, VS Code, Windsurf, Zed, Gemini CLI and the rest.
It listed its tools to an anonymous client, so reading the tool list needs no key. Individual tools can still ask for credentials in their own inputs.
Publisher: ai.theaggregate. Endpoint host: theaggregate.ai. ARD identifier: urn:air:ai.theaggregate:mcp:the-aggregate.
First indexed 31 August 2026. The tools are exactly what the server returned; nothing on this page
is inferred from its name. Ranking in Neuronto search is never for sale.