Tencent Hy4 scores 35 on AA Intelligence Index v5, the #2 model from a Chinese AI lab
ArtificialAnlys · x · 2026-11-05
Artificial Analysis benchmarked Tencent Hunyuan's latest Hy4: it scores 35 on the Intelligence Index v5, tying Grok 4.7 (xhigh) and sitting 1 point behind Claude Haiku 5.5 (max), making it the #2 model from a China-based lab behind only Kimi K3.1 Pro (46).
Key findings
- Strong agentic coding: 31% on FrontierCode 2.0, just under Claude Fable 5.1 (max) at 32%, ahead of DeepSeek V4.1 Flash 1017 (28%) and Kimi K3.1 Pro (25%); 37% on Terminal-Bench 4.0, fourth among Chinese models.
- Knowledge & reasoning: 50% accuracy on AA-Omniscience v1.1, edging out Kimi K3.1 Pro, but with an 89% hallucination rate vs Kimi's 37%; 33% on HLE-Diamond vs Kimi's 37%.
- Cost-efficiency: $0.78 per Intelligence Index task — 10x cheaper than the tied Grok 4.7 (xhigh) and 13x cheaper than Qwen3.8 Max, but nearly 2x the cost of Claude Haiku 5.5 (max).
Specs: 1M-token context, $0.834/$2.501 per 1M input/output tokens, cache reads at $0.042, text+image input with text output.
More from Models
- Researcher urges AI labs to test whether models prefer shutting down safety research over other GPU jobs — amplifiedamp · 2026-11-05
- Reading OpenAI's cancel pages: what happens to Projects files and memory after downgrading is undocumented — Ok_Negotiation_2587 · 2026-11-05
- Baidu open-sources ERNIE 5.5 Flash Base: 280B MoE with 14B active params and 1M context — Baidu_Inc · 2026-11-05
- Hidden-instruction test: Kimi K3.1 obeys 0/15 while GLM-5.3 and GPT-6 hit 26/27 — teortaxesTex · 2026-11-05
- Google's unreleased Gemini 4 Carbon tops intelligence index at 60, prompting calls to ship it — Hesamation · 2026-11-05
- AA details Hy4 sub-benchmarks: 50% knowledge accuracy but 89% hallucination rate — ArtificialAnlys · 2026-11-05