Who leads the WMT'23 benchmark?

GPT-4 currently leads WMT'23 with a score of 84.1 on Comet.

What is the state-of-the-art score on WMT'23?

The state-of-the-art result on WMT'23 is 84.1 (Comet), achieved by GPT-4 as of 2023.

How many models are tracked on WMT'23?

Codesota tracks 4 models on WMT'23.

When was the WMT'23 leaderboard last updated?

The WMT'23 leaderboard on Codesota includes results through 2023.

Codesota · Benchmark · WMT'23Home/Leaderboards/Language & Knowledge/Machine Translation/WMT'23

Unknown

WMT'23.

Name: WMT'23 Benchmark Results
Creator: Unknown
Published: 2023-01-01
License: https://creativecommons.org/licenses/by/4.0/

State-of-the-art machine translation evaluation from WMT 2023 shared task

Paper ↗Leaderboard ↓

§ 01 · Leaderboard

Results by metric.

Only 4 models on this benchmark

Help build the community leaderboard — submit your model results.

Found a wrong score or missing run?

Use row edits to send a sourced correction into moderation.

Add / edit result ↗Report issue ↗

Comet

Comet is the reported evaluation metric for WMT'23. 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 Cometverifiedpapervendorcommunityunverified

Rank	Model	Trust	Score	Year	Links	Fix
01	GPT-4 GPT-4 COMET-22 score on WMT23 en→de test set. From WMT23 General MT findings.	verified	84.1	2023	Source ↗	Looks wrong?
02	Google Translate Google Translate (ONLINE-B) COMET-22 on WMT23 en→de. Consistently top-tier online system.	verified	83.8	2023	Source ↗	Looks wrong?
03	DeepL DeepL (ONLINE-W) COMET-22 on WMT23 en→de. From WMT23 findings paper.	verified	83.5	2023	Source ↗	Looks wrong?
04	NLLB-3.3B NLLB-3.3B COMET-22 on WMT23 en→de. Open-source strong baseline.	verified	81.6	2023	Source ↗	Looks wrong?

§ 04 · Submit a result

Add to the leaderboard.

← Back to Machine Translation