---
type: "raw_signal"
id: "benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f"
slug: "benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f"
title: "Benchmark grades mergeability using rubrics from 20+ open-source maintainers , not just unit tests."
description: "Cognition has introduced FrontierCode, a benchmark intended to measure whether coding models produce changes maintainers would consider mergeable, not merely code that passes tests."
retrieval_nugget: "Cognition has introduced FrontierCode, a benchmark intended to measure whether coding models produce changes maintainers would consider mergeable, not merely code that passes tests."
observed_at: "2026-10-12"
record_date: "2026-10-12"
date_kind: "observed_at"
why_it_matters: "Cognition has introduced FrontierCode, a benchmark intended to measure whether coding models produce changes maintainers would consider mergeable, not merely code that passes tests. The company says the benchmark was built with maintainers of 36 open-source repositories, with more than 40 hours invested per task. Its criteria cover correctness, test quality, scope discipline, style, and repository"
novelty: "notable"
verification_level: "source-linked"
signal_type: "release"
source_platform: "cognition.com"
topics: []
entities: ["Cognition","FrontierCode","Claude Opus 4.8","SWE-Bench Pro","Kimi K2.6"]
related_patterns: []
source_url: "https://cognition.com/blog/frontier-code"
source_urls: ["https://cognition.com/blog/frontier-code"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f"
status: "published"
visuals: []
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"]}
routes: {"html":"https://newruntime.com/signals/benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f/","markdown":"https://newruntime.com/signals/benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f.md","json":"https://newruntime.com/signals/benchmark-grades-mergeability-using-rubrics-from-20-open-cfc01f.json"}
---

# Benchmark grades mergeability using rubrics from 20+ open-source maintainers , not just unit tests.

## Retrieval answer

Cognition has introduced FrontierCode, a benchmark intended to measure whether coding models produce changes maintainers would consider mergeable, not merely code that passes tests.

Cognition has introduced FrontierCode, a benchmark intended to measure whether coding models produce changes maintainers would consider mergeable, not merely code that passes tests. The company says the benchmark was built with maintainers of 36 open-source repositories, with more than 40 hours invested per task. Its criteria cover correctness, test quality, scope discipline, style, and repository-specific standards. Cognition describes 150 tasks across Extended, Main, and Diamond subsets. It reports Claude Opus 4.8 scores of 13.4% on Diamond, 34.3% on Main, and 51.8% on Extended. These are results from the benchmark’s creator; the supplied source does not provide an independent replication.

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)