Model
Hy3
Tencent Hy reasoning model for coding, instruction following, and agent tasks
5 providersreleased open weights
$0.066
best input /M · NanoGPT
$0.26
best output /M
256K
context window
64K
max output
Compare providers (sorted by listed input price)
| Provider | Input /M | Output /M | Cache read /M | Cache write /M | Context | Capabilities | Status | Updated |
|---|---|---|---|---|---|---|---|---|
| NanoGPT tencent/hy3 | $0.066 | $0.26 | $0.029 | — | 262K | reasoningtoolsstructuredno vision | Jul 6, 2026 | |
| OpenRouter tencent/hy3 | $0.132 | $0.528 | $0.033 | — | 262K | reasoningtoolsstructuredno vision | Jul 6, 2026 | |
| Vercel AI Gateway tencent/hy3 | $0.132 | $0.528 | $0.033 | — | 256K | reasoningtoolsno structuredno vision | Jul 6, 2026 | |
| Kilo Gateway tencent/hy3 | $0.14 | $0.58 | $0.035 | — | 262K | reasoningtoolsstructuredno vision | Jul 6, 2026 | |
| CrossModel tencent/hy3 | $0.16 | $0.64 | $0.04 | $0.16 | 262K | reasoningtoolsstructuredno vision | Jul 6, 2026 |
Price badge
https://models.sutraworks.ai/badge/tencent/hy3.svgLive SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.
Recent changes
Changelog →- repricedvia kilo · Aug 21, 2026
- input: $0.14 → $0.132
- output: $0.58 → $0.528
- cache_read: $0.035 → $0.033
Benchmarks
- SWE-Bench Verified ↗(resolved)report ↗78
Weights
Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.