Tencent Hunyuan Hy4 Flash hits 5.74 on RuneBench at ~$0.03 per run, cheapest on Pareto frontier
mariofilhoml · x · 2026-11-17
Third-party tester mariofilhoml benchmarked Tencent Hunyuan Hy4 Flash (high reasoning, run via Pi): 5.74 on RuneBench skills-30m at roughly $0.03 per run, making it the cheapest point on the cost-performance Pareto frontier — well below DeepSeek Flash ($0.123), GPT-6.1 ($1.1–$1.7), and GPT-6 Astra hi ($15.26 / 7.26).
Caveats: pricing is currently discounted, and the official RuneBench results come from @maxbittker.
More from Models
- OpenAI's flurry of blog posts sparks speculation of a model release today — imjustnewatai · 2026-11-17
- Sub-9GB open model bake-off: Qwen3.5-9B most balanced, IFM K2-7B tops math and coding — parepeg · 2026-11-17
- Polymarket claims GPT-6 Astra helped physicist crack 50-year particle physics problem — Polymarket · 2026-11-17
- Sam Altman says live: "We didn't have AGI, now we do" — imjustnewatai · 2026-11-17
- User says OpenAI refused $570 refund for auto-renewal with no explanation — OverLordGoldDragon · 2026-11-17
- Gemini spots the trick in a handwritten integral troll that stumps ChatGPT — vedantmisra · 2026-11-17