---
format: "aidr-story-markdown/v1"
id: "4947038be6a7bddf999e63767c935d25d8c99eb0123bce5e219790b76551d206"
canonical_url: "https://aidr.today/4947038b?lang=en"
title: "Tencent Hy4 AI Model Ranks No. 1 on CWE-Bench Security Test"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-13T09:57:32.000Z"
category: "Models"
topics: ["llm","safety"]
source_urls: ["https://marketbrief.now/ai/tencent-hy4-ai-model-ranks-no-1-on-cwe-bench-security-test-9039e1ce","https://huggingnews.com/ai/tencent-hy4-ai-model-ranks-no-1-on-cwe-bench-security-test-9039e1ce"]
summary: "Tencent's Hy4 AI model scored 76% on the CWE-Bench-AA test for finding and fixing software vulnerabilities. This score surpassed DeepSeek V4.1 Flash 1017 by 6 points. The Hy4 model costs $0.92 per task and uses 64k output tokens per task."
---

# Tencent Hy4 AI Model Ranks No\. 1 on CWE\-Bench Security Test

> [Open the canonical story](<https://aidr.today/4947038b?lang=en>)

**Published:** 2026-11-13T09:57:32.000Z
**Category:** Models
**Topics:** llm, safety

## Summary

Tencent's Hy4 AI model scored 76% on the CWE\-Bench\-AA test for finding and fixing software vulnerabilities\. This score surpassed DeepSeek V4\.1 Flash 1017 by 6 points\. The Hy4 model costs $0\.92 per task and uses 64k output tokens per task\.

## Sources

- [Story source](<https://marketbrief.now/ai/tencent-hy4-ai-model-ranks-no-1-on-cwe-bench-security-test-9039e1ce>)
- [Story source](<https://huggingnews.com/ai/tencent-hy4-ai-model-ranks-no-1-on-cwe-bench-security-test-9039e1ce>)