hy3 vs hy4-preview
| hy3 | hy4-preview | |
|---|---|---|
| Providers | Tencent Hunyuan | Tencent Hunyuan |
| Input $/M | $0.15 | $0.92 |
| Output $/M | $0.62 | $2.77 |
| Output multiplier | ×4.00 | ×3.00 |
| Context window | 256K tokens | 1M tokens |
| Intelligence index | 25.3 | — |
| Coding index | 58.8 | — |
What the numbers mean
hy3 is roughly 6.0× cheaper on input than hy4-preview. If your workload passes the quality bar on the cheaper model, that gap compounds across every request; route to hy4-preview deliberately rather than by default.
Tags: hy3 — Chat, Reasoning, Coding · hy4-preview — Chat, Reasoning, Coding
Code
from openai import OpenAI
client = OpenAI(base_url="https://starseaapi.com/v1", api_key="sk-...")
# hy3
client.chat.completions.create(model="hy3", messages=[{"role":"user","content":"Hi"}])
# hy4-preview
client.chat.completions.create(model="hy4-preview", messages=[{"role": "user", "content": "Hi"}])