all the models — AI benchmark observatory
← Models

EXAONE 4.5 33B

LG AI Research · proprietary · exaone-4.5-33b

EXAONE 4.5 is the first open-weight vision-language model from LG AI Research. It is a dense 33B multimodal model (31.7B language + 1.29B vision encoder) built by adding a dedicated visual encoder to the EXAONE 4.0 framework. It supports a 262,144-token context window, hybrid reasoning (thinking on by default), and strong performance on document understanding and Korean contextual reasoning while retaining EXAONE language capabilities. License: EXAONE AI Model License Agreement 1.2 - NC. Knowledge cutoff: December 2024.

AIME 2025AIME 2026IFEvalAI2DMathVista-MiMMLU-Pro

Benchmark scores

BenchmarkScore
AIME 20250.93
AIME 20260.93
IFEval0.90
AI2D0.89
MathVista-Mini0.85
MMLU-Pro0.83
GPQA0.81
MMMU0.79
MMMU-Pro0.69

Pricing

  • No provider pricing.

AA metrics

No Artificial Analysis link yet.

Arena Elo

  • No Arena snapshot linked.
EXAONE 4.5 33B Benchmarks · all the models