Project AURIONAI Use-Cases Research & Innovation
INRIA
tabular

TabICL v2

INRIA soda team tabular in-context learner; rank 4 on TabArena.

INRIAINRIAFrance
Chart of TabICLv2 average rank against number of samples, from the official repository.
Levelmodel
Classtabular
Parametersnot in the notes
Contextnot in the notes
Released2026
Opennessopen source
Runs onsingle datacentre GPU
Statusactive

Benchmark Results

maintainer-run
TabArena All Taskstabular
defaultmaintainer-run
1,598
8001,3001,800

Score in Elo; the axis starts at 800, as Elo has no zero. Higher is better.

rank 4 of 59, All Tasks, no imputation, snapshot 2026-05-06, 59 entries6 May 2026 benchmarklist.com, benchmarklist.com
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

Lineage

successor
TabICL to TabICL v2: successor, medium confidenceTabICLTabICLINRIA, 2025TabICL v2TabICL v2INRIA, 2026
INRIATabICL is the parent (successor)medium confidencenotes
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 2 models. Confidence and evidence are the notes' own.

R&D Topics

Tabular foundation modelsproduced by

Identity

Developer
INRIA
Family
not in the notes
Released
2026 sourcesnotes
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
dense transformer In-context transformer
Total parameters
– sources
Active parameters
–
Checkpoint parameters
–
Context length
not in the notes
Input
table
Output
table
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open source
Licence
not in the notes sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
single datacentre GPU
Footprint
Median train about 4.0 s
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Industry Use Cases

No industry use case links this model.

Sources

  1. TabArena All Tasks Benchmark Scores & AI Model Leaderboard | BenchmarkList benchmarklist.com https://benchmarklist.com/benchmarks/tabarena_all/
  2. The state of Tabular Foundation Models (2026) mindfulmodeler.substack.com https://mindfulmodeler.substack.com/p/the-state-of-tabular-foundation-models

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "TabICL v2", https://aurion.semai.id/models/inria-tabicl-v2/, records of October 2026.