Hy4 preview
Tencent · Open weights · Released Aug 28, 2026 · tencent/hy4-preview
Best $0.83 in / $2.50 out via Makora · fp8 · 131K ctx at this price (headline 1.05M)
1.05M contextToolsReasoningStructured outputPrompt cachingOpen weights
Available from 5 providersUSD / MTOK · PRICES 19 MINUTES AGO · SPEED 4 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | UPTIME |
|---|---|---|---|---|---|---|---|---|
| MAMakora | $0.83 | $2.50 | $0.041 | 131K | fp8 | — | — | 100.0% |
| $0.834 | $2.50 | $0.042 | 1.05M | fp8 | ✕ | ✕ | 100.0% | |
| $0.834 | $2.50 | $0.042 | 1.05M | fp8 | 1.8s | 41 tok/s | 100.0% | |
| $0.834 | $2.50 | $0.042 | 1M | fp8 | ✕ | ✕ | 99.8% | |
| $0.834 | $2.50 | $0.042 | 1.05M | fp8 | 1.5s | 69 tok/s | 100.0% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured via OpenRouter routing · ✕ = probed, endpoint returned no measurable stream · methodology v1.1 · measured 4 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Pricing variants · USD / MTok in / out
STANDARD
$0.83 / $2.50
per MTok in / out
CACHE READ $0.041
Recent changesRSS ↗
Oct 13, 2026endpoint added · makora4 → 5
Sep 29, 2026endpoint added · deepinfra3 → 4
Sep 23, 2026endpoint added · novita2 → 3
Sep 23, 2026endpoint added · siliconflow1 → 2
Aug 29, 2026model added
About
Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...
Specifications
Context window1,048,576
Max output64,000
Modalities intext
Modalities outtext
LicenseOpen weights
ReleasedAug 28, 2026
TokenizerOther
WeightsHugging Face ↗
Related
Embed badge
[](https://modelindex.ai/models/tencent/hy4-preview)Sources
OpenRouter API19 minutes ago
AGGREGATORHugging Face19 minutes ago
COMMUNITYModelIndex benchmarks8 days ago
MEASUREDVerified 19 minutes ago · list prices, not negotiated rates.