Hy3 preview
Tencent released 2026-04-22
$0.285 per million tokens, balanced
DeepSeek V4 Flash 0731 is both better and cheaper.
It scores +9.6 higher and costs 63% less ($0.105/M against $0.285/M) on this workload — and it does everything this model does.
Every price dimension
| Input | $0.180/M |
|---|---|
| Output | $0.600/M |
| Cached input | $0.060/M |
Independent scores
| Intelligence | 42.2 |
|---|---|
| Coding | 58.8 |
| Agentic | 31.4 |
Sources are listed separately rather than averaged. Across the 62 models both have scored they correlate at r = 0.806 — close agreement overall, but four models rank very differently between them, and a blended score would hide exactly those.
Capability
| Context window | 262K |
|---|---|
| Input modes | text |
| Tool use | yes |
| Reasoning | optional |
Provenance
| Price source | openrouter.ai |
|---|---|
| Fetched | 2026-08-24 |
| Quality data | verified |
Hy3 preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels across disabled, low, and high modes, allowing it to...