Who leads the TACRED benchmark?

LUKE currently leads TACRED with a score of 72.70 on f1.

What is the state-of-the-art score on TACRED?

The state-of-the-art result on TACRED is 72.70 (f1), achieved by LUKE as of 2026.

How many models are tracked on TACRED?

Codesota tracks 3 models on TACRED.

When was the TACRED leaderboard last updated?

The TACRED leaderboard on Codesota includes results through 2026.

Codesota · Knowledge Base · Relation Extraction · TACREDTasks/Knowledge Base/Relation Extraction

Relation Extraction · benchmark dataset · 2017 · EN

TAC Relation Extraction Dataset.

Name: TAC Relation Extraction Dataset Benchmark Results
Creator: Codesota
Published: 2026-01-01
License: https://creativecommons.org/licenses/by/4.0/

Large supervised relation extraction dataset from TAC KBP, 106K examples covering 41 relation types. Standard sentence-level RE benchmark evaluated by Micro-F1 (excluding no_relation).

Paper ↗Submit a result ↵

§ 01 · Leaderboard

Best published scores.

3 results indexed across 1 metric. Shaded row marks current SOTA; ties broken by submission date.

Primary: f1 · higher is better

f1· primary

3 rows

#	Model	Org	Submitted	Paper / code	f1
01	LUKEOSS	Studio Ousia	Apr 2026	editorial	72.70
02	KEPLEROSS	Tsinghua	Apr 2026	editorial	71.70
03	SpanBERTOSS	Facebook AI	Apr 2026	editorial	70.80

Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.

§ 03 · Progress

1 steps
of state of the art.

Each row below marks a model that broke the previous record on f1. Intermediate submissions are kept in the leaderboard above; only SOTA-setting entries are re-listed here.

Higher scores win. Each subsequent entry improved upon the previous best.

SOTA line · f1

Apr 5, 2026LUKEStudio Ousia72.70

Fig 3 · SOTA-setting models only. 1 entries span Apr 2026 → Apr 2026.

§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide

What a submission needs

01A public checkpoint or API endpoint
02A reproduction script with frozen commit + seed
03Declared evaluation environment (Python, deps)
04One row per metric declared by this dataset
05A contact so we can follow up on discrepancies

TAC Relation Extraction Dataset.

Best published scores.

1 stepsof state of the art.

Have a score that beatsthis table?

1 steps
of state of the art.

Have a score that beats
this table?