Primary use stated or inferred cautiously from official documentation.
T
TENCENT / hy4
Tencent Hy4
Stable Hy4 version available on Tencent TokenHub since November 4, 2026 with the identifier hy4. It adds image understanding and strengthens tool use and long tasks. It is distinct from Hy4 preview. Weight publication is announced for later; the Preview license and parameter counts are not inherited.
01 / SUMMARY
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
02 / TECHNICAL SHEET
Limits and integration
| API ID | hy4 |
|---|---|
| Model type | Multimodal · Code · Text and reasoning |
| Access model | Paid proprietary |
| License | Stable-version weights and license not yet verified |
| Deployment | Hosted API |
| Release | 04 NOV 2026 |
| Knowledge cutoff | Not published |
| Entrance | Text · Image |
| Exit | Text |
| Context window | 1M tokens; maximum 960K input on TokenHub |
| maximum output | 64K tokens on TokenHub |
| Reasoning | Reasoning enabled by default |
| Published tools | Features |
| Structured Outings | Documented structured output |
| Batch processing | Not published |
| Prompt cache | Not published |
| Fine-tuning | Not published |
| Verified platforms | Tencent TokenHub · OpenRouter |
Screenshot and chart analysis, programming and agents with tools. TokenHub documents compatibility with Chat Completions, Responses and Messages. Computer actions require a suitable environment and permissions; the model does not control devices on its own.
The 960K input and 64K output limits apply to TokenHub. Other channels may configure different limits. Productivity evaluations are internal and do not guarantee reliability in finance or autonomous actions. Stable-version weights have not been verified as downloadable.
03 / CAPABILITIES
What it can do
Orientation
Screenshot and chart analysis, programming and agents with tools. TokenHub documents compatibility with Chat Completions, Responses and Messages. Computer actions require a suitable environment and permissions; the model does not control devices on its own.
Context
1M tokens; maximum 960K input on TokenHub · 64K tokens on TokenHub
Tools and integration
Features
Access
Tencent TokenHub
Input modalities
TextImage
04 / PRICES
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | 6 CNY / 1M tokens before discount | Standard API rate |
| Cached input | 0.3 CNY / 1M tokens before discount | Reading reused prefixes |
| Cache write or storage | Not published | The condition varies by provider |
| Standard output | 18 CNY / 1M tokens before discount | May include reasoning tokens |
| Batch input | Not published | Asynchronous processing |
| Batch output | Not published | Asynchronous processing |
TokenHub lists base rates of 6 CNY input, 0.3 CNY cached input and 18 CNY output per million tokens. A 50% discount applies until December 31, 2026 at 23:59:59 Beijing time. Prices are not converted to USD using an assumed exchange rate. OpenRouter rates are independent.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
05 / EVALUATIONS
How to read the results
“
Tencent reports an average score of 83.18 and a 96.15% pass rate in a WorkBuddy evaluation with twelve scenarios. These are internal results for the product system, not an independent benchmark of the isolated model. Inferama has not reproduced the evaluation and does not equate it with general coding tests.
| Benchmark | Result | Metric | Source |
|---|---|---|---|
| Hy4 WorkBuddy internal pass rate | 96,15 % | Pass rate | View source ↗ |
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
06 / VERSIONS
Chronology
Tencent Hy4
Version added to Inferama's verified catalog.
07 / SOURCES