benchmarkheaven.com WebMCP Tools, Schemas & Agent Capabilities

benchmarkheaven.com

live Unverified (seed import) AI & Agents 6 tools imperative implementation

Where every benchmark result and price comes from, how the adjusted cost per task is calculated, and how publishers can ask for removal.

First seen 2026-11-05 · Last seen 2026-11-05 · Source: webmcp.com

Answer 5 · Act 1 · Transact 0 · Score: Not yet scored

Tool inspector

list_models

answer imperative

List systems on the current JevBench leaderboard (Jev-class decision models) with axis scores (intelligence, calibration, speed, cost; 0-100, higher is better), Capability score/rank/eligibility, Composite score/rank, listing flags, price per 1,000 decisions and median latency. From v1.7.0 (method v1.7) a decision over 10 s counts as failed: rows also give the measured p95, the number of decisions over 10 s, the method and the v1.6 composite. Read-only.

Page: /about

Input schema

Property	Type	Required	Description
`board`	string	no	"open" = open-weights board (/jev-models), "api" = hosted API board (/jev-models/api), "all" = both. The Jev reference row is on both boards.
`eligible_only`	boolean	no	Only systems eligible for the official Capability ranking (ranked and within the cost and median-latency caps).
`sort_by`	string	no	Score to sort by, descending (higher is better on every axis). Missing scores sort last; the key breaks ties.
`limit`	integer	no	—
`cursor`	string	no	next_cursor from the previous page.

Raw JSON schema

{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "board": {
      "type": "string",
      "enum": [
        "all",
        "open",
        "api"
      ],
      "default": "all",
      "description": "\"open\" = open-weights board (/jev-models), \"api\" = hosted API board (/jev-models/api), \"all\" = both. The Jev reference row is on both boards."
    },
    "eligible_only": {
      "type": "boolean",
      "default": false,
      "description": "Only systems eligible for the official Capability ranking (ranked and within the cost and median-latency caps)."
    },
    "sort_by": {
      "type": "string",
      "enum": [
        "capability",
        "composite",
        "intelligence",
        "calibration",
        "speed",
        "cost"
      ],
      "default": "capability",
      "description": "Score to sort by, descending (higher is better on every axis). Missing scores sort last; the key breaks ties."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 50,
      "default": 20
    },
    "cursor": {
      "type": "string",
      "pattern": "^[0-9]{1,4}$",
      "description": "next_cursor from the previous page."
    }
  }
}