Home/Models
Models/tencent

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.

tencent/hy4-preview
Organization
tencent
Family
Hy
Providers
15
Context
1,024,000
Output limit
64,000
Knowledge
—
Release
2026-08-28
Updated
2026-08-28
Weights
Open
Input
text
Output
text
Capabilities
Tools, Reasoning, Temperature

Providers

Compare every cataloged offer for this model, including its provider, route, token limits, price, and capabilities.

19 offers · USD per million tokens
ProviderModel IDContextOutputInput / output · 1MReasoningToolsStructuredDetails
AIHubMix
hy4-preview
1,024,00064,000$0.845 / $2.535YesYesYes
View
CrossModel
tencent/hy4-preview
1,048,57665,536$0.96 / $2.88YesYes—
View
Deep Infra
tencent/Hy4-preview
1,048,57664,000$0.834 / $2.501YesYesYes
View
DevPass (LLM Gateway)
hy4-preview
1,000,00064,000$0.834 / $2.501YesYes—
View
Eden AI
tencent/hy4-preview
1,000,00064,000$0.834 / $2.501YesNoNo
View
Eden AI
deepinfra/tencent/Hy4-preview
1,048,57664,000$0.834 / $2.501YesYesYes
View
Hugging Face
tencent/Hy4-preview
1,000,00064,000$0.834 / $2.501YesYesYes
View
Kilo Gateway
tencent/hy4-preview
1,048,57664,000$0.834 / $2.501YesYesYes
View
LLM Gateway
tencent/hy4-preview
1,000,00065,536$0.834 / $2.501YesYesYes
View
NanoGPT
tencent/hy4-preview
1,048,57664,000$0.834 / $2.501YesYesYes
View
OpenCode Go
hy4-preview
1,024,00064,000$0.834 / $2.501YesYes—
View
DeepInfravia OpenRouter
tencent/hy4-preview
fp8		1,048,576131,072$0.834 / $2.501YesYesYes
View
Novitavia OpenRouter
tencent/hy4-preview
fp8		1,000,00064,000$0.834 / $2.501YesYesYes
View
SiliconFlowvia OpenRouter
tencent/hy4-preview
fp8		1,048,576262,144$0.834 / $2.501YesYesYes
View
Tencentvia OpenRouter
tencent/hy4-preview
fp8		1,048,57664,000$0.834 / $2.501YesYesYes
View
Tencent TokenHub
hy4-preview
1,024,00064,000$0.834 / $2.501YesYes—
View
Tencent Token Plan
hy4-preview
1,024,00064,000$0.834 / $2.501YesYes—
View
Vancine
hy4-preview
1,024,00064,000$0.67 / $2.00YesYes—
View
Vercel AI Gateway
tencent/hy4-preview
1,024,00064,000$0.834 / $2.501YesYes—
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.

USD per million tokens
Input from$0.67Per 1M input tokens
Output from$2.00Per 1M output tokens
Available offers19Across 15 providers

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.

Endpoint price comparisonLatest observed snapshot · 4 lowest-priced configurations
InputOutput
DeepInfradeepinfra/fp8
$0.834$2.501
Novitanovita/fp8
$0.834$2.501
SiliconFlowsiliconflow/fp8
$0.834$2.501
Tencenttencent/fp8
$0.834$2.501
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
EndpointInput / 1MOutput / 1MCache read / 1MCache 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.

Latest 30-minute observation

Time to first token · milliseconds

Latency distributionRolling 30-minute snapshot · 4 endpoints
P50P75P90P99
DeepInfradeepinfra/fp8
1,691.5–5,023.13 ms
Novitanovita/fp8
4,822–10,842.8 ms
SiliconFlowsiliconflow/fp8
1,588–3,938.36 ms
Tencenttencent/fp8
3,961.5–12,654.49 ms
All latency percentiles 4 endpoints
EndpointP50P75P90P99
DeepInfradeepinfra/fp81,691.5 ms1,998 ms3,125.3 ms5,023.13 ms
Novitanovita/fp84,822 ms5,797 ms7,097.8 ms10,842.8 ms
SiliconFlowsiliconflow/fp81,588 ms1,874 ms2,396 ms3,938.36 ms
Tencenttencent/fp83,961.5 ms4,649.75 ms5,690.9 ms12,654.49 ms

Output throughput · tokens per second

Throughput distributionRolling 30-minute snapshot · 4 endpoints
P50P75P90P99
DeepInfradeepinfra/fp8
26–30.92 t/s
Novitanovita/fp8
29–81 t/s
SiliconFlowsiliconflow/fp8
50–84 t/s
Tencenttencent/fp8
48–95 t/s
All throughput percentiles 4 endpoints
EndpointP50P75P90P99
DeepInfradeepinfra/fp826 t/s30 t/s30.2 t/s30.92 t/s
Novitanovita/fp829 t/s37 t/s49.1 t/s81 t/s
SiliconFlowsiliconflow/fp850 t/s58 t/s67 t/s84 t/s
Tencenttencent/fp848 t/s62 t/s75 t/s95 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.

Successful requests by window
Availability by endpointCurrent overlapping windows · 4 endpoints
Endpoint5 minutes30 minutes24 hours
DeepInfradeepinfra/fp8
Novitanovita/fp8
SiliconFlowsiliconflow/fp8
Tencenttencent/fp8

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.

External evaluations

Sources: Artificial Analysis and Design Arena, via OpenRouter. Scores retain their original scales and are not OpenAI Suite ratings.

Detailed evaluations

Evaluation scoresAccuracy or item F1 · higher is better
tau bench verified airlineTencent: Hy4 preview
70.667%
gpqa diamondTencent: Hy4 preview
89.057%
EvaluationConfigurationSourceResultsDetails
tau bench verified airlineTencent: Hy4 previewtencent/hy4-preview-20260827OpenRouter70.667%
Details
gpqa diamondTencent: Hy4 previewtencent/hy4-preview-20260827OpenRouter89.057%
Details

Sources: Artificial Analysis, Design Arena and OpenRouter. Evaluation configurations and scales differ; results are not interchangeable. Updated Oct 4, 2026.

Apps & session costs

Compare observed median session costs for applications using this model, grouped by conversation length.

30-day sample · through Sep 27, 2026
ApplicationModel routeTurnsMedian cost / sessionDetails
Claude Codetencent/hy4-preview-2026082750+ turns
$1.196602
Details
Hermes Agenttencent/hy4-preview-202608272–9 turns
$0.02523
Details
Hermes Agenttencent/hy4-preview-2026082750+ turns
$0.874254
Details
Claude Codetencent/hy4-preview-202608272–9 turns
$0.03894
Details
Hermes Agenttencent/hy4-preview-2026082710–49 turns
$0.171622
Details
Claude Codetencent/hy4-preview-2026082710–49 turns
$0.259739
Details
Hermes Agenttencent/hy4-preview-202608271 turn
$0.006238
Details

Source: OpenRouter, as of Oct 2, 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.

Last 90 complete UTC days
Daily token usage62,256,170,399,888 reported tokens · 37 days with observations
Jul 6, 2026Oct 3, 2026
Daily observations · 37 days
Date (UTC)Reported tokensReported routes
Oct 3, 2026766,158,622,8421
Oct 2, 2026665,223,669,4461
Oct 1, 2026672,539,410,3751
Sep 30, 20261,048,211,814,9861
Sep 29, 20261,178,735,652,4191
Sep 28, 20261,211,198,041,2761
Sep 27, 20261,001,910,248,5691
Sep 26, 2026874,031,239,2911
Sep 25, 2026826,828,502,1131
Sep 24, 20261,342,529,612,0721
Sep 23, 20261,629,400,121,5241
Sep 22, 20262,084,298,005,0871
Sep 21, 20261,883,026,566,6931
Sep 20, 20261,971,310,447,6121
Sep 19, 20261,411,341,498,0401
Sep 18, 20262,026,493,739,2671
Sep 17, 20261,983,748,730,0111
Sep 16, 20261,581,018,240,1171
Sep 15, 20261,818,993,397,2601
Sep 14, 20261,724,288,980,4431
Sep 13, 20261,379,623,656,2821
Sep 12, 20261,493,844,286,1041
Sep 11, 20261,628,503,208,2381
Sep 10, 20262,330,317,015,0781
Sep 9, 20263,254,378,053,5201
Sep 8, 20263,405,147,893,4311
Sep 7, 20263,324,245,026,6461
Sep 6, 20261,766,332,794,8911
Sep 5, 20262,043,385,142,1471
Sep 4, 20262,954,935,990,4381
Sep 3, 20262,989,073,301,5171
Sep 2, 20262,265,371,409,4481
Sep 1, 20261,714,603,276,2421
Aug 31, 2026939,373,453,4081
Aug 30, 20261,163,086,114,3821
Aug 29, 20261,538,875,055,6581
Aug 28, 2026363,788,183,0151

Source: OpenRouter (openrouter.ai/rankings), as of Oct 4, 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.

Current route specifications
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
Base URLhttps://openrouter.ai/api/v1
Model IDtencent/hy4-preview
Version IDtencent/hy4-preview-20260827
Weights IDtencent/Hy4-preview

Reasoning

Required
No
Enabled by default
Yes
Default effort
high
Supported efforts
high, low, none

Supported parameters

  • frequency_penalty
  • include_reasoning
  • logit_bias
  • max_completion_tokens
  • max_tokens
  • min_p
  • presence_penalty
  • reasoning
  • reasoning_effort
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_k
  • top_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 $0.67 per million input tokens and $2.00 per million output tokens. Prices vary by provider and configuration; see Pricing for details.

Which providers offer Hy4 preview?

There are 19 cataloged offers across 15 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: Yes. Structured output: Not reported. Support can vary by endpoint.

Catalog history · 2 recorded revisions

Metadata observations since this model was first cataloged. These are not model release versions.

  • · 4cc85ba86534
  • · 50a46117b0b7