Hy3 Preview
Hy3 Preview is Tencent's language model with a 262K context window and up to 236K output tokens, available from 2 providers, starting at $0.18 / 1M input and $0.6 / 1M output. A high-efficiency Mixture-of-Experts LLM from Tencent's Hunyuan family, designed for agentic workflows with configurable reasoning levels for production use.
| Benchmarks | |
|---|---|
| GPQA | #81 |
| HLE | #117 |
| IFBench | #108 |
| Time to First Token | #326 |
| SciCode | #147 |
| LCR | #191 |
| TerminalBench Hard | #80 |
| TAU2 | #50 |
| Output TPS | #558 |
Capabilities
Input1/5
Text✓
Image·
Audio·
Video·
PDF·
Output1/5
Text✓
Image·
Audio·
Video·
Embedding·
Capabilities3/13
Reasoning✓
Adaptive Reasoning·
Function Calling✓
Parallel Function Calling·
Structured Outputs·
Native JSON Schema·
Web Search·
URL Context·
Computer Use·
Code Execution·
File Search·
Prompt Caching✓
Assistant Prefill·
Pricing by Provider
US Dollar ($)
Per 1M tokens
| Provider | Standard | ||
|---|---|---|---|
| Input $ / 1M | Output $ / 1M | Cache Read $ / 1M | |
| $0.18 | $0.6 | $0.06 | |
| $0.18 | $0.6 | $0.06 | |
Cost Calculator
US Dollar ($)
Preset:
Cheapest Instances to Run It
Cloud GPU instances that can host Hy3 Preview, ranked by cheapest on-demand price. The model needs about 717 GB of GPU memory at FP16 precision (estimated from its parameter count), so treat the fit as guidance rather than a guarantee.
All clouds
FP16 (full precision)
US Dollar ($)
Instance | Cloud | GPU | VRAM | Price | Cheapest region | |
|---|---|---|---|---|---|---|
| g7e.48xlarge |  AWS | 8× RTX PRO Server 6000 | 768 GB | $33.14/hr | ||
| g4-standard-384 |  GCP | 8× nvidia-rtx-pro-6000 | 768 GB | $36.00/hr | ||
| Standard_ND96isr_MI300X_v5 |  Azure | 8× AMD Instinct MI300X | 1536 GB | $48.00/hr | ||
Versions
| Version | Released | Context | Input / 1M | Output / 1M | Status |
|---|---|---|---|---|---|
| Hy 3 | 262K | — | — | Available | |
| Hy3 Preview | 262K | $0.180 | $0.600 | Current | |
| Hunyuan A13B Instruct | 131K | $0.140 | $0.570 | Available |