Hunyuan HY3 Preview

Tencent / Hunyuan
Model overview

hy3-preview

The open-weight HY3 preview retained for reproducible research and migration comparisons after the GA release.

general LLMreasoningcodinglong contextagentevaluationcode generationlocal inferenceenterprise access
Representative version

hy3-preview

CompanyTencent / Hunyuan
Release2026-04-23
Release typePreview release
Parameters295B / 21B active
ArchitectureNot disclosed
Context262,144
Input modalitiestext / code
Output modalitiestext / code / tool_call
Tool use / function callingConfirmed: Tool use
Structured output / reasoningConfirmed: Reasoning
Vision / audio / videoVision: Not disclosed; Audio: Not disclosed; Video: Not disclosed
Open weights / LicenseWeights available / Tencent Hunyuan community license (license source pending verification)
tencent-hy3-preview-current · License source pending verification
Vendor APIAvailable
Tencent Hunyuan HY3 Preview repositorytencent-hy3-preview-current · 2026-04-23
OpenRouterOpenRouter listed
PricingOpenRouter · 2026-08-11: $0.063 / $0.21
input / output, per 1M tokens
BenchmarkNo reliable public source
Public signalsNo reliable public source
Use casesresearch evaluation / Migration testing / local inference / code generation
Claim evidenceRelease event / Model specifications / Weight checkpointTencent Hunyuan HY3 Preview repositoryhy3-preview · 2026-04-23 · Hunyuan HY3 Preview · hy3-preview · hy3-preview
OpenRouter context / capabilities / pricingOpenRouter public models APItencent/hy3-preview · 2026-08-11T06:28:09.895Z
OpenRouter

Pricing / context / availability

OpenRouterOpenRouter listed
OpenRouter listingtencent/hy3-preview
Mapped versionhy3-preview
tencent-hy3-preview-current
PricingOpenRouter · 2026-08-11: $0.063 / $0.21
input / output, per 1M tokens
Context262144
Speed/LatencyNo reliable public source
Snapshot time2026-08-11T06:28:09.895Z
SourceView source
Verified limitations

Usage boundaries

  • Superseded by the HY3 GA model; uses a separate community license.
Missing public information

Public information not yet available

No reliable benchmark sourceNo public model-card activity countersNo linkable model-specific reviewSpeed/latency not disclosed