The AI arena is free today

Open Superagent

Hy4 preview vs Phi-3.5-vision-instruct

Hy4 preview leads the LLM Stats Score 51.3 to -3.1.

Tencent · Microsoft · Updated for 2026

Which is better?

Hy4 preview leads the overall LLM Stats Score 51.3 to -3.1, ranking #13 overall.

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

Choose Hy4 preview

  • overall performance matters — it scores 51.3 and ranks #13 on LLM Stats
  • your work emphasizes reasoning — it leads those capability indexes
  • you want the most recent training data — it shipped Aug 2026

Choose Phi-3.5-vision-instruct

  • you are already invested in the Microsoft ecosystem

At a glance

The differences that matter most.

Core performance indexes
51.3
#13
-3.1
#331
51.0
#10
-4.5
#328
Cost, coverage & limits
Benchmark wins
Input price
— / M
— / M
Output price
— / M
— / M
Context window

Individual benchmarks

32 reported for Hy4 preview · 9 for Phi-3.5-vision-instruct

No common benchmarks found

Hy4 preview and Phi-3.5-vision-instructdon't have any common benchmark datasets to compare. They may have been evaluated on different testing suites.

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

765.8B diff

Hy4 preview has 765.8B more parameters than Phi-3.5-vision-instruct, making it 18233.3% larger.

Tencent
Hy4 preview
770.0Bparameters
Microsoft
Phi-3.5-vision-instruct
4.2Bparameters
770.0B
Hy4 preview
4.2B
Phi-3.5-vision-instruct

Input capabilities

Documented input modalities across available providers

Phi-3.5-vision-instruct supports multimodal inputs, whereas Hy4 preview does not.

Phi-3.5-vision-instruct can handle both text and other forms of data like images, making it suitable for multimodal applications.

Hy4 preview

Text
Images
Audio
Video

Phi-3.5-vision-instruct

Text
Images
Audio
Video

License

Usage and distribution terms

Hy4 preview is licensed under Apache 2.0, while Phi-3.5-vision-instruct uses MIT.

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

Hy4 preview

Apache 2.0

Open weights

Phi-3.5-vision-instruct

MIT

Open weights

Release Timeline

When each model was launched

Hy4 preview was released on 2026-08-28, while Phi-3.5-vision-instruct was released on 2024-08-23.

Hy4 preview is 25 months newer than Phi-3.5-vision-instruct.

Hy4 preview

Aug 28, 2026

5 days ago

2.0yr newer
Phi-3.5-vision-instruct

Aug 23, 2024

2.0 years ago

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against Hy4 preview and Phi-3.5-vision-instruct side-by-side, then vote on the output you prefer.

Hy4 preview
✓ Preferred
Phi-3.5-vision-instruct
Open in Playground

FAQ

Common questions about Hy4 preview vs Phi-3.5-vision-instruct.

Which is better, Hy4 preview or Phi-3.5-vision-instruct?

Hy4 preview leads the LLM Stats Score 51.3 to -3.1. Hy4 preview is made by Tencent and Phi-3.5-vision-instruct is made by Microsoft. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Hy4 preview compare to Phi-3.5-vision-instruct in benchmarks?

Hy4 preview scores GPQA: 92.3%, Terminal-Bench 2.1: 85.4%, WideSearch: 83.9%, MCP Atlas: 83.7%, SWE-bench Multilingual: 82.9%. Phi-3.5-vision-instruct scores ScienceQA: 91.3%, POPE: 86.1%, MMBench: 81.9%, ChartQA: 81.8%, AI2D: 78.1%.

What are the main differences between Hy4 preview and Phi-3.5-vision-instruct?

Key differences include LLM Stats Score (51.3 vs -3.1), multimodal support (no vs yes), licensing (Apache 2.0 vs MIT). See the full comparison above for benchmark-by-benchmark results.

Who makes Hy4 preview and Phi-3.5-vision-instruct?

Hy4 preview is developed by Tencent and Phi-3.5-vision-instruct is developed by Microsoft.