---
type: "raw_signal"
id: "search-capability-leaderboard-24-models-ranked-on-web-a50c9a"
slug: "search-capability-leaderboard-24-models-ranked-on-web-a50c9a"
title: "Search Capability Leaderboard: 24 models ranked on web search"
description: "Parallel has launched the Search Capability Leaderboard, a public comparison of 24 open- and closed-source models used with web search."
retrieval_nugget: "Parallel has launched the Search Capability Leaderboard, a public comparison of 24 open- and closed-source models used with web search."
observed_at: "2026-09-25"
record_date: "2026-09-25"
date_kind: "observed_at"
why_it_matters: "Source: https://parallel.ai/blog/search-leaderboard"
novelty: "notable"
verification_level: "source-linked"
signal_type: "research"
source_platform: "parallel.ai"
topics: []
entities: []
related_patterns: []
source_url: "https://parallel.ai/blog/search-leaderboard"
source_urls: ["https://parallel.ai/blog/search-leaderboard"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:search-capability-leaderboard-24-models-ranked-on-web-a50c9a"
status: "published"
visuals: []
routes: {"html":"https://newruntime.com/signals/search-capability-leaderboard-24-models-ranked-on-web-a50c9a/","markdown":"https://newruntime.com/signals/search-capability-leaderboard-24-models-ranked-on-web-a50c9a.md","json":"https://newruntime.com/signals/search-capability-leaderboard-24-models-ranked-on-web-a50c9a.json"}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
---

# Search Capability Leaderboard: 24 models ranked on web search

## Retrieval answer

Parallel has launched the Search Capability Leaderboard, a public comparison of 24 open- and closed-source models used with web search.

Parallel has launched the Search Capability Leaderboard, a public comparison of 24 open- and closed-source models used with web search. It reports a Search Intelligence Score based on three existing public benchmark suites and a Search Efficiency ranking based on cost among models meeting or exceeding median accuracy.

In its September 15, 2026 snapshot, Parallel reports GPT-6 (Astra) as the highest-scoring model on Search Intelligence at 70.3 and $301 per 1,000 tasks. It reports GPT-5.6 Luna as the Search Efficiency leader at 55.0 and $23 per 1,000 tasks, while DeepSeek V4.1 Flash gained 45 points when paired with search.

Source: https://parallel.ai/blog/search-leaderboard

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)