The AI arena is free today

Open Superagent

Gemini 2.5 Pro Preview 06-05 vs Step3-VL-10B

Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B are closely matched at 29.9 and 26.1 on the LLM Stats Score.

Google · StepFun · Updated for 2026

Which is better?

Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B are closely matched on the overall LLM Stats Score at 29.9 and 26.1.

In the 2 individual benchmarks reported for both models, Gemini 2.5 Pro Preview 06-05 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 Gemini 2.5 Pro Preview 06-05

  • you value its reported benchmark strengths — it wins 2 of 2 exact shared results

Choose Step3-VL-10B

  • you want the most recent training data — it shipped Jan 2026
  • you need open weights you can self-host or fine-tune

At a glance

The differences that matter most.

Gemini 2.5 Pro Preview 06-05
Google
Step3-VL-10B
StepFun
Core performance indexes
LLM Stats Score
29.9
#116
26.1
#143
Reasoning
28.3
#120
25.9
#140
Cost, coverage & limits
Benchmark wins
2 of 2
0 of 2
Input price
$1.25 / M
— / M
Output price
$10.00 / M
— / M
Context window
1,048,576
—

Capability indexes

Additional strengths measured across groups of related public benchmarks

3 shared
Index
Gemini 2.5 Pro Preview 06-05
Step3-VL-10B
Math
22.2#134
23.9#121
Vision
18.9#73
17.1#81
Multimodal
25.0#48
20.9#63
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

13 reported for Gemini 2.5 Pro Preview 06-05 · 6 for Step3-VL-10B

2 shared

Gemini 2.5 Pro Preview 06-05 outperforms in 2 benchmarks (AIME 2025, MMMU), while Step3-VL-10B is better at 0 benchmarks.

Gemini 2.5 Pro Preview 06-05 significantly outperforms across most benchmarks.

Sat Sep 05 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Context Window

Maximum input and output token capacity

Only Gemini 2.5 Pro Preview 06-05 specifies input context (1,048,576 tokens). Only Gemini 2.5 Pro Preview 06-05 specifies output context (65,535 tokens).

Google
Gemini 2.5 Pro Preview 06-05
Input1,048,576 tokens
Output65,535 tokens
StepFun
Step3-VL-10B
Input- tokens
Output- tokens
Sat Sep 05 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Both Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B support multimodal inputs.

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

Gemini 2.5 Pro Preview 06-05

Text
Images
Audio
Video

Step3-VL-10B

Text
Images
Audio
Video

License

Usage and distribution terms

Gemini 2.5 Pro Preview 06-05 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.

Gemini 2.5 Pro Preview 06-05

Proprietary

Closed source

Step3-VL-10B

Apache 2.0

Open weights

Release Timeline

When each model was launched

Gemini 2.5 Pro Preview 06-05 was released on 2025-06-05, while Step3-VL-10B was released on 2026-01-15.

Step3-VL-10B is 7 months newer than Gemini 2.5 Pro Preview 06-05.

Gemini 2.5 Pro Preview 06-05

Jun 5, 2025

1.3 years ago

Step3-VL-10B

Jan 15, 2026

7 months ago

7mo newer

Knowledge Cutoff

When training data ends

Gemini 2.5 Pro Preview 06-05 has a documented knowledge cutoff of 2025-01-31, while Step3-VL-10B's cutoff date is not specified.

We can confirm Gemini 2.5 Pro Preview 06-05's training data extends to 2025-01-31, but cannot make a direct comparison without Step3-VL-10B's cutoff date.

Gemini 2.5 Pro Preview 06-05

Jan 2025

Step3-VL-10B

—

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Judge for yourself.

Run your own prompts against Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B side-by-side, then vote on the output you prefer.

Gemini 2.5 Pro Preview 06-05
✓ Preferred
Step3-VL-10B
Open in Playground

FAQ

Common questions about Gemini 2.5 Pro Preview 06-05 vs Step3-VL-10B.

Which is better, Gemini 2.5 Pro Preview 06-05 or Step3-VL-10B?

Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B are closely matched on the LLM Stats Score at 29.9 and 26.1. Gemini 2.5 Pro Preview 06-05 is made by Google 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 Gemini 2.5 Pro Preview 06-05 compare to Step3-VL-10B in benchmarks?

Gemini 2.5 Pro Preview 06-05 scores Global-MMLU-Lite: 89.2%, AIME 2025: 88.0%, FACTS Grounding: 87.8%, GPQA: 86.4%, VideoMMMU: 83.6%. Step3-VL-10B scores MMBench: 91.8%, AIME 2025: 87.7%, MathVista: 84.0%, MMMU: 78.1%, MathVision: 70.8%.

What are the context window sizes for Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B?

Gemini 2.5 Pro Preview 06-05 supports 1.0M tokens and Step3-VL-10B supports an unknown number of tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B?

Key differences include LLM Stats Score (29.9 vs 26.1), licensing (Proprietary vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes Gemini 2.5 Pro Preview 06-05 and Step3-VL-10B?

Gemini 2.5 Pro Preview 06-05 is developed by Google and Step3-VL-10B is developed by StepFun.

Related comparisons