Hy3
Hy3 is a large language model made by Tencent, added to the catalog on 6 Jul 2026. It has a context window of 262,144 tokens and costs $0.14 per 1M input tokens and $0.58 per 1M output tokens. It accepts text and produces text. Typical uses: reviewing long documents, multi-step text reasoning and tool-assisted workflows.
Price checked ยท source: OpenRouter
What Hy3 is
Tencent Hy3 is a text-in, text-out model from Tencent, described for reasoning, agentic workflows, and real-world production use. It accepts text and produces text. Its listed capabilities include reasoning, tools, structured outputs, and response-format controls. This combination is relevant to tasks where a text response needs to follow a requested format or where a surrounding application supplies tools for an interaction. The catalog does not specify which tools or output schemas are available, so those details depend on the integration.
Hy3 has a context length of 262,144 tokens and a maximum output length of 128,000 tokens. The context window can hold extensive text in a single interaction, which can be useful when a task refers to long documents, multiple passages, or detailed instructions. Its output allowance supports generating longer text responses when a task calls for them. These limits describe capacity; they do not guarantee a particular result or response quality.
The model is suited to text-based reasoning, tool-supported workflows, and structured text generation. Example applications include analyzing lengthy written material, producing responses in a specified format, and handling tasks divided into reasoning steps. Its listed input and output modalities are text only. The available facts do not list image or audio input or output, nor do they identify specific integrations, domains, or specialized applications. People evaluating Hy3 can consider its text capabilities, context and output limits, and listed interaction features against the requirements of their workflows.
Key features
- Accepts text and produces text
- Has a 262,144-token context length
- Supports reasoning and tool use
- Supports structured outputs and response-format controls
Typical uses
- Reviewing long documents
- Multi-step text reasoning
- Tool-assisted workflows
- Generating structured responses
Hy3 pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $0.14 |
| Output, per 1M tokens | $0.58 |
| Cached input read, per 1M tokens | $0.035 |
The price last changed on 22 Nov 2026. 3 price changes have been recorded since Hy3 was added.
| Date | Input /1M | Output /1M | Change |
|---|---|---|---|
| $0.14 | $0.58 | Price raised | |
| $0.0825 | $0.33 | Price lowered | |
| $0.132 | $0.528 | Price raised | |
| $0.0825 | $0.33 | Added at this price |
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0002 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.0040 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.0054 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.022 |
Hy3 specifications
| Model id | tencent/hy3 |
|---|---|
| Version | hy3-20260706 |
| Context window | 262,144 tokensAbout 196,608 English words |
| Max output | 235,929 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 6 Jul 2026 |
| Input types | text |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Hy3 yet.
Compare Hy3
Side by side with other models: prices, context window, what each accepts and supports, and the same four tasks priced on both.