Answers one row per model for the benchmark named — what our own harness…

Answers one row per model for the benchmark named — what our own harness measured, beside what the vendor claims, and the gap between them.

GET /v1/benchmark/leaderboard

Benchmark → · All API references → · Six flows, four surfaces →


Address	`https://api.hanzo.ai/v1/benchmark/leaderboard`
Method	`GET`
Operation	`get_benchmark_leaderboard`
Auth	`Authorization: Bearer $HANZO_API_KEY`

Answers one row per model for the benchmark named — what our own harness measured, beside what the vendor claims, and the gap between them.

The gap is the point of the arena; provider-reported claims have run materially hot against one standardized harness.

The two planes are NEVER blended, and that is the rule to read the rows by: a model we have measured but no vendor has claimed for shows published null, a model with only a claim shows measured null, and gap exists only where both do.

n is coverage and is not decoration: two measured numbers taken over different item counts are not comparable, so read the row's n before reading its accuracy.

Request

1 field.

Field	In	Type	Required	Description
`benchmark`	query	string	—	Benchmark is the catalog id to read, defaulting to gpqa_diamond.

Response

Status	Body	Meaning
`200`	leaderboard	ok

200 body — 15 fields.

Field	In	Type	Always	Description
`benchmark`	body	string	—	Benchmark is the catalog id these rows are about.
`rows`	body	LeaderRow[]	—	Rows is one per model, ordered by measured accuracy descending.
`rows[].ciHigh`	body	number	—	CIHigh is the upper bound of that interval.
`rows[].ciLow`	body	number	—	CILow and CIHigh are the 95% Wilson interval on Measured, in percent.
`rows[].claims`	body	integer	—	Claims is how many independent claims exist for this model on this benchmark.
`rows[].gap`	body	number	—	published − measured (the arena signal)
`rows[].mean`	body	number	—	Mean is the unweighted average of every claim, which answers a different question from Published: what the field says on average, rather than what the vendor…
`rows[].measured`	body	number	—	hanzo-measured accuracy % (nil if unrun)
`rows[].measuredAt`	body	string (date-time)	—	MeasuredAt is when the run behind Measured was recorded.
`rows[].model`	body	string	—	the model this row scores
`rows[].n`	body	integer	—	coverage — NEVER compare across different n
`rows[].protocol`	body	string	—	how the vendor scored their claim: single-attempt, pass@k or agentic
`rows[].published`	body	number	—	provider-claimed % (nil if none)
`rows[].run`	body	string	—	Run names the measurement Measured came from, and MeasuredAt is when it ran.
`rows[].spread`	body	number	—	Spread is the distance between the highest and lowest of them, nil when there is only one.

Failure carries the platform error shape — see Errors.

Examples

hanzo benchmark leaderboard

import { Configuration, BenchmarkApi } from 'hanzoai';

const api = new BenchmarkApi(new Configuration({ accessToken: process.env.HANZO_API_KEY }));
const { data } = await api.getBenchmarkLeaderboard();

from hanzoai.cloud import ApiClient, Configuration
from hanzoai.cloud.api import BenchmarkApi

client = ApiClient(Configuration(access_token=os.environ["HANZO_API_KEY"]))
result = BenchmarkApi(client).get_benchmark_leaderboard()

cfg := cloud.NewConfiguration()
cfg.AddDefaultHeader("Authorization", "Bearer "+os.Getenv("HANZO_API_KEY"))
client := cloud.NewAPIClient(cfg)

resp, _, err := client.BenchmarkAPI.GetBenchmarkLeaderboard(context.Background()).Execute()
if err != nil {
	return err
}

use hanzo_cloud::apis::{configuration::Configuration, benchmark_api};

let mut cfg = Configuration::new();
cfg.bearer_access_token = std::env::var("HANZO_API_KEY").ok();

let result = benchmark_api::get_benchmark_leaderboard(&cfg, Default::default()).await?;

import ai.hanzo.cloud.ApiClient;
import ai.hanzo.cloud.api.BenchmarkApi;

ApiClient client = new ApiClient();
client.setRequestInterceptor(b -> b.header("Authorization", "Bearer " + System.getenv("HANZO_API_KEY")));

var result = new BenchmarkApi(client).getBenchmarkLeaderboard();

curl https://api.hanzo.ai/v1/benchmark/leaderboard \
  -H "Authorization: Bearer $HANZO_API_KEY"

Tool benchmark, op get_benchmark_leaderboard — POST the JSON-RPC envelope to https://api.hanzo.ai/v1/mcp.

curl -X POST https://api.hanzo.ai/v1/mcp \
  -H "Authorization: Bearer $HANZO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
       "jsonrpc": "2.0",
       "id": 1,
       "method": "tools/call",
       "params": {
         "name": "benchmark",
         "arguments": {
           "op": "get_benchmark_leaderboard",
           "input": {}
         }
       }
     }'

Benchmark API · All Hanzo APIs · Interactive reference

Answers one row per model for the benchmark named — what our own harness…

Request

Response

Examples

On this page