Author's Description
Hy4 flash is a cost-efficient variant of [Hy4](https://openrouter.ai/tencent/hy4) from Tencent, served with roughly a quarter of the flagship's parameter count for lower latency and cost. It targets code assistance, agent...
Key Specifications
Supported Parameters
This model supports the following parameters:
Features
This model supports the following features:
Performance Summary
Tencent's Hy4 flash delivers strong all-around capabilities while maintaining cost-effective pricing, ranking in the 71st percentile for affordability. Operationally, it demonstrates exceptional reliability, achieving a 99% success rate with virtually no technical failures across benchmarks. The model's standout strength lies in complex reasoning, where it earned an impressive 98.0% accuracy (88th percentile) at very low cost, alongside excellent multi-step instruction following at 84.0% accuracy (87th percentile). It also demonstrates solid domain competence in structured tasks, posting 93.0% accuracy in both mathematics and multi-language coding evaluations. Furthermore, its massive 1M-token context window makes it highly adaptable for long-context agentic workflows. The primary trade-off is generation speed; Hy4 flash shows moderate speed performance overall (32nd percentile), exhibiting notably longer execution times on comprehensive coding evaluations. Despite these moderate speeds, Hy4 flash offers an attractive combination of high reasoning fidelity, rock-solid reliability, and budget-friendly economics, making it a dependable choice for developers building cost-sensitive code assistants and analytical agents.
Model Pricing
Current Pricing
| Feature | Price (per 1M tokens) |
|---|---|
| Prompt | $0.035 |
| Completion | $0.141 |
| Input Cache Read | $0.0015 |
Price History
Available Endpoints
| Provider | Endpoint Name | Context Length | Pricing (Input) | Pricing (Output) |
|---|---|---|---|---|
|
Tencent
|
Tencent | tencent/hy4-flash-20261116 | 1M | $0.035 / 1M tokens | $0.141 / 1M tokens |
Benchmark Results
| Benchmark | Category | Reasoning | Strategy | Free | Executions | Accuracy | Cost | Duration |
|---|
Other Models by tencent
|
|
Released | Params | Context |
|
Speed | Ability | Cost |
|---|---|---|---|---|---|---|---|
| Tencent: Hy4 | Nov 03, 2026 | — | 1M |
Image input
Text input
Text output
|
★ | ★★★ | $$$$$ |
| Tencent: Hy4 preview | Aug 27, 2026 | ~770B | 1M |
Text input
Text output
|
★ | ★★★★ | $$$$$ |
| Tencent: Hy-MT2-1.8B | Aug 20, 2026 | 8B | 8K |
Text input
Text output
|
★★★★★ | ★ | $$ |
| Tencent: Hy-MT2-30B-A3B | Aug 20, 2026 | 30B | 8K |
Text input
Text output
|
★★★★ | ★★ | $$ |
| Tencent: Hy-MT2-7B | Aug 19, 2026 | 7B | 8K |
Text input
Text output
|
★★★ | ★ | $$ |
| Tencent: Hy3 | Jul 06, 2026 | ~295B | 262K |
Text input
Text output
|
★ | ★★ | $$ |
| Tencent: Hy3 preview | Apr 22, 2026 | — | 262K |
Text input
Text output
|
★ | ★★★★ | $$$$ |
| Tencent: Hunyuan A13B Instruct Unavailable | Jul 08, 2025 | 13B | 131K |
Text input
Text output
|
★★ | ★ | $$ |