Who leads the CNN/DailyMail benchmark?

BRIO currently leads CNN/DailyMail with a score of 47.78 on rouge-1.

What is the state-of-the-art score on CNN/DailyMail?

The state-of-the-art result on CNN/DailyMail is 47.78 (rouge-1), achieved by BRIO as of 2024.

How many models are tracked on CNN/DailyMail?

Codesota tracks 7 models on CNN/DailyMail across 3 metrics.

When was the CNN/DailyMail leaderboard last updated?

The CNN/DailyMail leaderboard on Codesota includes results through 2024, with the earliest tracked result from 2019.

Codesota · Benchmark · CNN/DailyMailHome/Leaderboards/Language & Knowledge/Text Summarization/CNN/DailyMail

Unknown

CNN/DailyMail.

Name: CNN/DailyMail Benchmark Results
Creator: Unknown
Published: 2019-01-01
License: https://creativecommons.org/licenses/by/4.0/

300K news articles with multi-sentence summaries. Standard benchmark for abstractive summarization.

Paper ↗Leaderboard ↓

§ 01 · SOTA history

Year over year.

§ 02 · Leaderboard

Results by metric.

Found a wrong score or missing run?

Use row edits to send a sourced correction into moderation.

Add / edit result ↗Report issue ↗

rouge-1

Rouge 1 is the reported evaluation metric for CNN/DailyMail. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for rouge-1verifiedpapervendorcommunityunverified

Rank	Model	Trust	Score	Year	Links	Fix
01	BRIO BRIO (BART-large backbone). Source: Table 1, arxiv:2203.16804.	verified	47.78	2022	Paper ↗	Looks wrong?
02	GPT-4o GPT-4o zero-shot summarization.	verified	46.3	2023	Paper ↗Source ↗	Looks wrong?
03	Gemini 1.5 Pro Gemini 1.5 Pro zero-shot. Source: Gemini 1.5 technical report.	verified	45.8	2024	Paper ↗	Looks wrong?
04	Llama 3.1 405B Llama 3.1 405B Instruct zero-shot summarization. Source: Llama 3 paper.	verified	45.1	2024	Paper ↗	Looks wrong?
05	Qwen2 72B Qwen2 72B Instruct zero-shot. Source: Qwen2 technical report.	verified	44.7	2024	Paper ↗	Looks wrong?
06	PEGASUS-Large PEGASUS-Large. Source: Table 1, arxiv:1912.08777.	verified	44.17	2019	Paper ↗	Looks wrong?

rouge-l

Rouge L is the reported evaluation metric for CNN/DailyMail. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for rouge-lverifiedpapervendorcommunityunverified

Rank	Model	Trust	Score	Year	Links	Fix
01	BRIO BRIO. Source: Table 1, arxiv:2203.16804.	verified	44.57	2022	Paper ↗	Looks wrong?
02	GPT-4o GPT-4o zero-shot summarization.	verified	43.4	2023	Paper ↗Source ↗	Looks wrong?
03	Gemini 1.5 Pro Gemini 1.5 Pro zero-shot. Source: Gemini 1.5 technical report.	verified	43	2024	Paper ↗	Looks wrong?
04	Llama 3.1 405B Llama 3.1 405B Instruct zero-shot summarization. Source: Llama 3 paper.	verified	42.3	2024	Paper ↗	Looks wrong?
05	Qwen2 72B Qwen2 72B Instruct. Source: Qwen2 technical report.	verified	41.8	2024	Paper ↗	Looks wrong?
06	PEGASUS-Large PEGASUS-Large. Source: Table 1, arxiv:1912.08777.	verified	41.11	2019	Paper ↗	Looks wrong?
07	BART	unverified	40.9	2019	Paper ↗Code ↗	Looks wrong?

rouge-2

Rouge 2 is the reported evaluation metric for CNN/DailyMail. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for rouge-2verifiedpapervendorcommunityunverified

Rank	Model	Trust	Score	Year	Links	Fix
01	BRIO BRIO. Source: Table 1, arxiv:2203.16804.	verified	23.55	2022	Paper ↗	Looks wrong?
02	GPT-4o GPT-4o zero-shot summarization.	verified	22.1	2023	Paper ↗Source ↗	Looks wrong?
03	PEGASUS-Large PEGASUS-Large. Source: Table 1, arxiv:1912.08777.	verified	21.47	2019	Paper ↗	Looks wrong?

§ 04 · Submit a result

Add to the leaderboard.

← Back to Text Summarization