Hy4 preview
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 require planning, context continuity, and sustained multi-step execution.
- Organization
- tencent
- Family
- Hy
- Providers
- 19
- Context
- 1,024,000
- Output limit
- 64,000
- Knowledge
- —
- Release
- 2026-08-28
- Updated
- —
- Weights
- Open
- Input
- text
- Output
- text
- Capabilities
- —
Providers
Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.
| Provider | Model ID | Context | Output | Input / output · 1M | Reasoning | Tools | Structured | Details |
|---|---|---|---|---|---|---|---|---|
| AIHubMix | hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| AI/ML API | tencent/hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| CommandCode | tencent/hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| CrossModel | tencent/hy4-preview | 1,048,576 | 65,536 | — / — | — | — | — | View |
| Deep Infra | tencent/Hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| DevPass (LLM Gateway) | hy4-preview | 1,000,000 | 64,000 | — / — | — | — | — | View |
| Eden AI | tencent/hy4-preview | 1,000,000 | 64,000 | — / — | — | — | — | View |
| Eden AI | deepinfra/tencent/Hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| Hugging Face | tencent/Hy4-preview | 1,000,000 | 64,000 | — / — | — | — | — | View |
| Kilo Gateway | tencent/hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| LLM Gateway | tencent/hy4-preview | 1,000,000 | 65,536 | — / — | — | — | — | View |
| NanoGPT | tencent/hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| Novita AI | tencent/hy4-preview | 1,000,000 | 64,000 | — / — | — | — | — | View |
| OpenCode Go | hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| DeepInfravia OpenRouter | tencent/hy4-preview fp8 | 1,048,576 | 131,072 | $0.834 / $2.501 | Yes | Yes | Yes | View |
| Novitavia OpenRouter | tencent/hy4-preview fp8 | 1,000,000 | 64,000 | $0.834 / $2.501 | Yes | Yes | Yes | View |
| SiliconFlowvia OpenRouter | tencent/hy4-preview fp8 | 1,048,576 | 262,144 | $0.834 / $2.501 | Yes | Yes | Yes | View |
| Tencentvia OpenRouter | tencent/hy4-preview fp8 | 1,048,576 | 64,000 | $0.834 / $2.501 | Yes | Yes | Yes | View |
| Tencent TokenHub | hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| Tencent Token Plan | hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| Vancine | hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
| Venice AI | hy4-preview | 1,048,576 | 64,000 | — / — | — | — | — | View |
| Vercel AI Gateway | tencent/hy4-preview | 1,024,000 | 64,000 | — / — | — | — | — | View |
Prices and limits apply to each offer. A dash means the value is not provided. OpenRouter offers show the hosting provider and routing channel separately.
Pricing
Compare token rates and cache charges across available routes and hosting configurations.
Lowest listed input and output prices across the offers above; they may belong to different providers. These are base rates, before conditional pricing, discounts or additional fees.
tencent/hy4-previewObserved - Input
- $0.834
- Output
- $2.501
- Cache read
- $0.042
Additional fees and pricing conditions
overrides · USD per token for token prices; time windows use UTC.
[
{
"prompt": "0.000000834",
"utc_end": 1600,
"utc_start": 0,
"completion": "0.000002501",
"input_cache_read": "0.000000042"
},
{
"prompt": "0.0000007506",
"utc_end": 0,
"utc_start": 1600,
"completion": "0.0000022509",
"input_cache_read": "0.0000000378"
}
]Hosting prices through OpenRouter All endpoints and cache rates
| Endpoint | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M |
|---|---|---|---|---|
| DeepInfradeepinfra/fp8 | $0.834 | $2.501 | $0.042 | — |
| Novitanovita/fp8 | $0.834 | $2.501 | $0.042 | — |
| SiliconFlowsiliconflow/fp8 | $0.834 | $2.501 | $0.042 | — |
| Tencenttencent/fp8 | $0.834 | $2.501 | $0.042 | — |
Each row is a hosting configuration. Open an offer in Providers for its full pricing conditions and additional fees. Endpoint observations may differ from the routing catalog's base price.
Performance
Review reported latency and output-throughput percentiles from the latest observed rolling window.
Time to first token · milliseconds
All latency percentiles 4 endpoints
| Endpoint | P50 | P75 | P90 | P99 |
|---|---|---|---|---|
| DeepInfradeepinfra/fp8 | 5,347 ms | 8,776 ms | 14,343.2 ms | 27,645.48 ms |
| Novitanovita/fp8 | 5,232 ms | 6,041 ms | 6,982.8 ms | 20,172.66 ms |
| SiliconFlowsiliconflow/fp8 | 1,813.5 ms | 2,216 ms | 3,361.1 ms | 23,320.26 ms |
| Tencenttencent/fp8 | 4,036 ms | 4,804.25 ms | 6,037.5 ms | 13,182.34 ms |
Output throughput · tokens per second
All throughput percentiles 4 endpoints
| Endpoint | P50 | P75 | P90 | P99 |
|---|---|---|---|---|
| DeepInfradeepinfra/fp8 | 11 t/s | 17 t/s | 24 t/s | 29.59 t/s |
| Novitanovita/fp8 | 31 t/s | 46 t/s | 64 t/s | 91.11 t/s |
| SiliconFlowsiliconflow/fp8 | 37 t/s | 43 t/s | 48 t/s | 56 t/s |
| Tencenttencent/fp8 | 47 t/s | 60 t/s | 73 t/s | 92.09 t/s |
Percentiles describe the observed request distribution. These are snapshots of a rolling window, not a historical time series.
Uptime
Compare successful request rates across the latest 5-minute, 30-minute, and 24-hour observations.
Reported by OpenRouter at the time of collection; rate-limited requests are excluded. A dash means no measurement was reported. This is not a live availability check.
Benchmarks
Review attributed evaluation results on each source's original scale.
Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.
Detailed evaluations
| Evaluation | Configuration | Source | Results | Details |
|---|---|---|---|---|
| gpqa diamond | Tencent: Hy4 previewtencent/hy4-preview-20260827 | OpenRouter | 90.236% | Details |
| tau bench verified airline | Tencent: Hy4 previewtencent/hy4-preview-20260827 | OpenRouter | 76.222% | Details |
Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 24, 2026.
Apps & session costs
Compare observed median session costs for applications using this model, grouped by conversation length.
| Application | Model route | Turns | Median cost / session | Details |
|---|---|---|---|---|
| Hermes Agent | tencent/hy4-preview-20260827 | 2–9 turns | $0.027266 | Details |
| Claude Code | tencent/hy4-preview-20260827 | 10–49 turns | $0.250315 | Details |
| Hermes Agent | tencent/hy4-preview-20260827 | 50+ turns | $0.814753 | Details |
| Hermes Agent | tencent/hy4-preview-20260827 | 10–49 turns | $0.182374 | Details |
Source: OpenRouter, as of Oct 20, 2026. Published under CC BY 4.0. These are observed costs per session, not token prices or a forecast for your workload. Apps are compared separately.
Activity
Daily prompt and completion tokens reported for this model's routes in OpenRouter's top 50. Missing days have no published value.
363,788,183,015 tokens
1,538,875,055,658 tokens
1,163,086,114,382 tokens
939,373,453,408 tokens
1,714,603,276,242 tokens
2,265,371,409,448 tokens
2,989,073,301,517 tokens
2,954,935,990,438 tokens
2,043,385,142,147 tokens
1,766,332,794,891 tokens
3,324,245,026,646 tokens
3,405,147,893,431 tokens
3,254,378,053,520 tokens
2,330,317,015,078 tokens
1,628,503,208,238 tokens
1,493,844,286,104 tokens
1,379,623,656,282 tokens
1,724,288,980,443 tokens
1,818,993,397,260 tokens
1,581,018,240,117 tokens
1,983,748,730,011 tokens
2,026,493,739,267 tokens
1,411,341,498,040 tokens
1,971,310,447,612 tokens
1,883,026,566,693 tokens
2,084,298,005,087 tokens
1,629,400,121,524 tokens
1,342,529,612,072 tokens
826,828,502,113 tokens
874,031,239,291 tokens
1,001,910,248,569 tokens
1,211,198,041,276 tokens
1,178,735,652,419 tokens
1,048,211,814,986 tokens
672,539,410,375 tokens
665,223,669,446 tokens
766,158,622,842 tokens
967,647,974,174 tokens
1,064,912,183,986 tokens
1,233,511,174,686 tokens
1,366,269,153,136 tokens
1,294,556,736,746 tokens
1,269,781,435,108 tokens
1,209,688,057,277 tokens
840,734,460,945 tokens
1,099,108,443,116 tokens
1,049,936,538,401 tokens
1,073,519,654,710 tokens
1,336,938,324,688 tokens
1,242,284,978,846 tokens
907,386,160,031 tokens
845,637,992,998 tokens
1,369,069,954,570 tokens
1,382,907,094,450 tokens
1,462,587,748,707 tokens
1,458,128,550,852 tokens
1,345,210,358,109 tokens
Daily observations · 57 days
| Date (UTC) | Reported tokens | Reported routes |
|---|---|---|
| Oct 23, 2026 | 1,345,210,358,109 | 1 |
| Oct 22, 2026 | 1,458,128,550,852 | 1 |
| Oct 21, 2026 | 1,462,587,748,707 | 1 |
| Oct 20, 2026 | 1,382,907,094,450 | 1 |
| Oct 19, 2026 | 1,369,069,954,570 | 1 |
| Oct 18, 2026 | 845,637,992,998 | 1 |
| Oct 17, 2026 | 907,386,160,031 | 1 |
| Oct 16, 2026 | 1,242,284,978,846 | 1 |
| Oct 15, 2026 | 1,336,938,324,688 | 1 |
| Oct 14, 2026 | 1,073,519,654,710 | 1 |
| Oct 13, 2026 | 1,049,936,538,401 | 1 |
| Oct 12, 2026 | 1,099,108,443,116 | 1 |
| Oct 11, 2026 | 840,734,460,945 | 1 |
| Oct 10, 2026 | 1,209,688,057,277 | 1 |
| Oct 9, 2026 | 1,269,781,435,108 | 1 |
| Oct 8, 2026 | 1,294,556,736,746 | 1 |
| Oct 7, 2026 | 1,366,269,153,136 | 1 |
| Oct 6, 2026 | 1,233,511,174,686 | 1 |
| Oct 5, 2026 | 1,064,912,183,986 | 1 |
| Oct 4, 2026 | 967,647,974,174 | 1 |
| Oct 3, 2026 | 766,158,622,842 | 1 |
| Oct 2, 2026 | 665,223,669,446 | 1 |
| Oct 1, 2026 | 672,539,410,375 | 1 |
| Sep 30, 2026 | 1,048,211,814,986 | 1 |
| Sep 29, 2026 | 1,178,735,652,419 | 1 |
| Sep 28, 2026 | 1,211,198,041,276 | 1 |
| Sep 27, 2026 | 1,001,910,248,569 | 1 |
| Sep 26, 2026 | 874,031,239,291 | 1 |
| Sep 25, 2026 | 826,828,502,113 | 1 |
| Sep 24, 2026 | 1,342,529,612,072 | 1 |
| Sep 23, 2026 | 1,629,400,121,524 | 1 |
| Sep 22, 2026 | 2,084,298,005,087 | 1 |
| Sep 21, 2026 | 1,883,026,566,693 | 1 |
| Sep 20, 2026 | 1,971,310,447,612 | 1 |
| Sep 19, 2026 | 1,411,341,498,040 | 1 |
| Sep 18, 2026 | 2,026,493,739,267 | 1 |
| Sep 17, 2026 | 1,983,748,730,011 | 1 |
| Sep 16, 2026 | 1,581,018,240,117 | 1 |
| Sep 15, 2026 | 1,818,993,397,260 | 1 |
| Sep 14, 2026 | 1,724,288,980,443 | 1 |
| Sep 13, 2026 | 1,379,623,656,282 | 1 |
| Sep 12, 2026 | 1,493,844,286,104 | 1 |
| Sep 11, 2026 | 1,628,503,208,238 | 1 |
| Sep 10, 2026 | 2,330,317,015,078 | 1 |
| Sep 9, 2026 | 3,254,378,053,520 | 1 |
| Sep 8, 2026 | 3,405,147,893,431 | 1 |
| Sep 7, 2026 | 3,324,245,026,646 | 1 |
| Sep 6, 2026 | 1,766,332,794,891 | 1 |
| Sep 5, 2026 | 2,043,385,142,147 | 1 |
| Sep 4, 2026 | 2,954,935,990,438 | 1 |
| Sep 3, 2026 | 2,989,073,301,517 | 1 |
| Sep 2, 2026 | 2,265,371,409,448 | 1 |
| Sep 1, 2026 | 1,714,603,276,242 | 1 |
| Aug 31, 2026 | 939,373,453,408 | 1 |
| Aug 30, 2026 | 1,163,086,114,382 | 1 |
| Aug 29, 2026 | 1,538,875,055,658 | 1 |
| Aug 28, 2026 | 363,788,183,015 | 1 |
Source: OpenRouter (openrouter.ai/rankings), as of Oct 24, 2026. CC BY 4.0. Totals include only individually published routes. Tokenizers differ by provider; the aggregated “other” category is never assigned to a model.
Technical details & API
Explore specifications, supported parameters, reasoning controls, and API identifiers for each OpenRouter route.
tencent/hy4-previewObserved - Context
- 1,048,576
- Maximum output
- 64,000
- Tokenizer
- Other
- Input
- text
- Output
- text
- Moderation
- No
- Added to OpenRouter
- 2026-08-28
https://openrouter.ai/api/v1tencent/hy4-previewtencent/hy4-preview-20260827tencent/Hy4-previewReasoning
- Required
- No
- Enabled by default
- Yes
- Default effort
- high
- Supported efforts
- high, low, none
Supported parameters
frequency_penaltyinclude_reasoninglogit_biasmax_completion_tokensmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_p
Frequently asked questions
Answers based on the model metadata and serving offers currently in the catalog.
How much does Hy4 preview cost?
Listed rates start at $0.834 per million input tokens and $2.501 per million output tokens. Prices vary by provider and configuration; see Pricing for details.
Which providers offer Hy4 preview?
There are 23 cataloged offers across 19 providers. Compare model IDs, prices and limits in Providers.
What is the context window?
The cataloged model context is 1,024,000 tokens. Each serving endpoint may apply a different limit.
Does it support tools and structured output?
Tool calling: Not reported. Structured output: Not reported. Support can vary by endpoint.
Catalog history · 3 recorded revisions
Metadata observations since this model was first cataloged. These are not model release versions.
- ·
9dc8b19fce8f - ·
4cc85ba86534 - ·
50a46117b0b7