Artificial Analysis scores Tencent's Hy4 35 on its Intelligence Index, #2 among China-based labs

Artificial Analysis says Tencent's Hy4 scores 35 on its Intelligence Index v5, level with Grok 4.7 (xhigh) and one point behind Claude Haiku 5.5 (max). It ranks Hy4 second among models from China-based labs, behind Kimi K3.1 Pro at 46.

According to Artificial Analysis:

• Coding: Hy4 scores 31% on FrontierCode 2.0 (Claude Fable 5.1 at max scores 32%), the top result from a China-based lab ahead of DeepSeek V4.1 Flash 1017 at 28% and Kimi K3.1 Pro at 25%. On Terminal-Bench 4.0 it scores 37%, fourth among China-based labs.
• Knowledge: 50% accuracy on AA-Omniscience v1.1, one point above Kimi K3.1 Pro, but an 89% hallucination rate against Kimi K3.1 Pro's 37%. On HLE-Diamond it scores 33% against Kimi K3.1 Pro's 37%.
• Cost: about $0.78 per Intelligence Index task, which it calls 10 times cheaper than Grok 4.7 (xhigh) and 13 times cheaper than Qwen3.8 Max (0902), and almost twice the cost of Claude Haiku 5.5 (max).
• Model details: 1 million-token context, text and image input with text output, priced at $0.834 per million input tokens, $2.501 per million output tokens and $0.042 per million cached input tokens.

These are Artificial Analysis's own measurements and the prices it lists.