The AI arena is free today

Open Superagent

EXAONE 4.5 33B vs Step3-VL-10B

EXAONE 4.5 33B and Step3-VL-10B are closely matched at 26.3 and 26.1 on the LLM Stats Score.

LG AI Research · StepFun · Updated for 2026

Which is better?

EXAONE 4.5 33B and Step3-VL-10B are closely matched on the overall LLM Stats Score at 26.3 and 26.1.

In the 2 individual benchmarks reported for both models, EXAONE 4.5 33B wins 2; this is a narrower head-to-head signal than the composite indexes.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose EXAONE 4.5 33B

  • you value its reported benchmark strengths — it wins 2 of 2 exact shared results
  • you want the most recent training data — it shipped Apr 2026

Choose Step3-VL-10B

  • you need open weights you can self-host or fine-tune

At a glance

The differences that matter most.

Core performance indexes
26.3
#141
26.1
#143
26.9
#134
25.9
#140
Cost, coverage & limits
Benchmark wins
2 of 2
0 of 2
Input price
— / M
— / M
Output price
— / M
— / M
Context window

Capability indexes

Additional strengths measured across groups of related public benchmarks

3 shared
Index
EXAONE 4.5 33B
Step3-VL-10B
29.1#85
23.9#121
14.1#91
17.1#81
17.0#75
20.9#63
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

15 reported for EXAONE 4.5 33B · 6 for Step3-VL-10B

2 shared

EXAONE 4.5 33B outperforms in 2 benchmarks (AIME 2025, MMMU), while Step3-VL-10B is better at 0 benchmarks.

EXAONE 4.5 33B significantly outperforms across most benchmarks.

Sat Sep 05 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

23.0B diff

EXAONE 4.5 33B has 23.0B more parameters than Step3-VL-10B, making it 230.0% larger.

LG AI Research
EXAONE 4.5 33B
33.0Bparameters
StepFun
Step3-VL-10B
10.0Bparameters
33.0B
EXAONE 4.5 33B
10.0B
Step3-VL-10B

Input capabilities

Documented input modalities across available providers

Both EXAONE 4.5 33B and Step3-VL-10B support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

EXAONE 4.5 33B

Text
Images
Audio
Video

Step3-VL-10B

Text
Images
Audio
Video

License

Usage and distribution terms

EXAONE 4.5 33B is licensed under a proprietary license, while Step3-VL-10B uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

EXAONE 4.5 33B

Proprietary

Closed source

Step3-VL-10B

Apache 2.0

Open weights

Release Timeline

When each model was launched

EXAONE 4.5 33B was released on 2026-04-09, while Step3-VL-10B was released on 2026-01-15.

EXAONE 4.5 33B is 3 months newer than Step3-VL-10B.

EXAONE 4.5 33B

Apr 9, 2026

4 months ago

2mo newer
Step3-VL-10B

Jan 15, 2026

7 months ago

Knowledge Cutoff

When training data ends

EXAONE 4.5 33B has a documented knowledge cutoff of 2024-12-01, while Step3-VL-10B's cutoff date is not specified.

We can confirm EXAONE 4.5 33B's training data extends to 2024-12-01, but cannot make a direct comparison without Step3-VL-10B's cutoff date.

EXAONE 4.5 33B

Dec 2024

Step3-VL-10B

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against EXAONE 4.5 33B and Step3-VL-10B side-by-side, then vote on the output you prefer.

EXAONE 4.5 33B
✓ Preferred
Step3-VL-10B
Open in Playground

FAQ

Common questions about EXAONE 4.5 33B vs Step3-VL-10B.

Which is better, EXAONE 4.5 33B or Step3-VL-10B?

EXAONE 4.5 33B and Step3-VL-10B are closely matched on the LLM Stats Score at 26.3 and 26.1. EXAONE 4.5 33B is made by LG AI Research and Step3-VL-10B is made by StepFun. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does EXAONE 4.5 33B compare to Step3-VL-10B in benchmarks?

EXAONE 4.5 33B scores AIME 2025: 92.9%, AIME 2026: 92.6%, IFEval: 89.6%, AI2D: 89.0%, MathVista-Mini: 85.0%. Step3-VL-10B scores MMBench: 91.8%, AIME 2025: 87.7%, MathVista: 84.0%, MMMU: 78.1%, MathVision: 70.8%.

What are the main differences between EXAONE 4.5 33B and Step3-VL-10B?

Key differences include LLM Stats Score (26.3 vs 26.1), licensing (Proprietary vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes EXAONE 4.5 33B and Step3-VL-10B?

EXAONE 4.5 33B is developed by LG AI Research and Step3-VL-10B is developed by StepFun.