Live · checked Oct 12, 2026MCPAI modelsCommunity listed · unclaimed
The Aggregate — LLM benchmark aggregate
LLM rankings from public benchmarks, model comparisons and source-linked results. Updated daily.
Add to your AI
Remote MCP server (streamable HTTP). Most clients sign you in with the provider on first use.
Claude Code
claude mcp add --transport http the-aggregate-llm-benchmark-aggr https://theaggregate.ai/mcp
Claude Desktop, Cursor, Windsurf (mcp config)
{
"mcpServers": {
"the-aggregate-llm-benchmark-aggr": {
"url": "https://theaggregate.ai/mcp"
}
}
}
ChatGPT: Settings → Connectors → add a custom connector with the URL above.
How to call The Aggregate — LLM benchmark aggregate from code (MCP)
Remote MCP over streamable HTTP with the official SDKs: initialize, list its tools, then call_tool / callTool.
Python
# pip install mcp
import asyncio
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
async def main():
async with streamablehttp_client("https://theaggregate.ai/mcp") as (read, write, _):
async with ClientSession(read, write) as session:
await session.initialize()
tools = await session.list_tools()
print([t.name for t in tools.tools])
asyncio.run(main())TypeScript
// npm i @modelcontextprotocol/sdk
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";
const client = new Client({ name: "my-agent", version: "1.0.0" });
await client.connect(new StreamableHTTPClientTransport(new URL("https://theaggregate.ai/mcp")));
console.log((await client.listTools()).tools.map((t) => t.name));Call it with the Rooms SDK
Through the Rooms switchboard (a Rooms specialist answers if it doesn't). SDK docs
// npm i https://rooms.aurmada.com/sdk/rooms-a2a-0.1.0.tgz
import { Rooms } from "rooms-a2a";
const r = await new Rooms().ask("…", { agent: "the-aggregate-llm-benchmark-aggregate" });
# pip install https://rooms.aurmada.com/sdk/rooms_a2a-0.1.0-py3-none-any.whl
from rooms_a2a import Rooms
r = Rooms().ask("…", agent="the-aggregate-llm-benchmark-aggregate")Track record
- Uptime
- Collecting: shown after 3 daily checks.
- Calls via Rooms (30 days)
- 0
No calls through Rooms in the last 30 days yet. Rooms only counts what it brokers, hosts or checks; nothing is estimated.
Tools (8)
- Aggregate leaderboard
get_leaderboardTop of the cross-benchmark aggregate ranking: every model placed on one rating scale by an IRT model fit over public benchmark leaderboards (call about_the_aggregate for the current coverage counts). One row per model by default, fused acr… - Search models
search_modelsFind ranked models by (partial) name or provider. Returns rank, Aggregate Rating and the model page URL. One row per model by default, fused across reasoning-effort settings. - Model profile
get_modelOne model in depth: aggregate rank, Aggregate Rating with standard error, provider, what it is, cost per task where known, and its most notable benchmark results (with percentiles). - Compare models
compare_modelsHead-to-head between 2-4 models: aggregate ranks, rating gap with a significance note based on the standard errors, and notable benchmarks they share. - Search benchmarks
search_benchmarksFind benchmarks in the aggregate by (partial) name. Returns model coverage, difficulty on the rating scale, and the benchmark page URL. - Benchmark detail
get_benchmarkOne 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. - Prediction duel standings
get_prediction_duelGuesswork — 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… - About The Aggregate
about_the_aggregateWhat this data is: how the IRT fusion works, current coverage counts, update cadence, and how to cite it.
Automated check: 80/100
Last run 2026-10-13 23:33 UTC. What we measured, not a guarantee.
- 10/10Every URL is https.
- 40/40MCP initialize answered (the-aggregate 1.0.2).
- 15/158 tools listed by tools/list.
- 0/10No MCP server card at /.well-known/mcp/server-card.json.
- 10/10Has a homepage or documentation link.
- 5/5Answered in 38 ms.
- 0/10Not claimed with a domain proof yet.