Skip to main content
AI Choice Engine

Tencent

Tencent Hy4 Preview

Tencent Hy4 Preview is a Tencent model with 1.05M context, $0.834 in / $2.501 out per million tokens, last verified October 25, 2026. Open weights: yes. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.

Tencent Hy4 Preview is an Apache-licensed open-weight text MoE with long native context and hosted TokenHub access.

Recorded profile verification October 25, 2026Individual provider fields may changeAccess: preview — Provider-labelled open-weight preview, also listed on Tencent Cloud TokenHub; service activation and regional endpoint credentials are required.

AI model specification details
SpecificationTencent Hy4 Preview
ProviderTencent
TierFrontier
Context window1.05M
Max outputNot verifiedUnverified
Input / 1M tokens$0.834
Output / 1M tokens$2.501
WeightsOpen
Parameters770B backbone total / 49B active MoE; separate 10B / 0.7B active MTP layer
Reasoning levelsnone, high
Modalitiestext
LicenseApache 2.0
API model idhy4-preview
ReleasedAugust 28, 2026

Pricing tiers: Tencent Cloud TokenHub lists $0.834/$2.501 per MTok, with $0.042 cache-hit input. The checkpoint config has 1,048,576 native context; the hosted guide advertises context/input/output as 1M/960K/64K. An exact integer output ceiling is unverified. App access and self-hosted costs are separate.

Best for

  • Open-weight frontier work
  • Long-context
  • Self-hosting at scale

Watch out

Preview checkpoint with acknowledged excessive reasoning and over-verification. TokenHub reasoning_effort accepts none/high and defaults high; do not inherit Hy3's low setting. The 163-expert/203-task blind comparison is Tencent-reported, not an independent benchmark rerun. Keep Apache notices when redistributing weights. The launch's two-week free WorkBuddy/CodeBuddy offer does not establish current free access.

Source receipts

Catalog figures for this model were checked against the following sources.

Common questions

Tencent Hy4 Preview

Answered from the verified figures on this page rather than general guidance.

How much does Tencent Hy4 Preview cost per million tokens?
Tencent Hy4 Preview lists base rates of $0.834 per million input tokens and $2.501 per million output tokens. Estimate both input and output token volume and any additional charges for your workload. Tencent Cloud TokenHub lists $0.834/$2.501 per MTok, with $0.042 cache-hit input. The checkpoint config has 1,048,576 native context; the hosted guide advertises context/input/output as 1M/960K/64K. An exact integer output ceiling is unverified. App access and self-hosted costs are separate.
What is Tencent Hy4 Preview's context window?
Tencent Hy4 Preview accepts about 1.05M tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.
What is Tencent Hy4 Preview best for?
Tencent Hy4 Preview is a frontier tier from Tencent. It suits open-weight frontier work, long-context, self-hosting at scale. Preview checkpoint with acknowledged excessive reasoning and over-verification. TokenHub reasoning_effort accepts none/high and defaults high; do not inherit Hy3's low setting. The 163-expert/203-task blind comparison is Tencent-reported, not an independent benchmark rerun. Keep Apache notices when redistributing weights. The launch's two-week free WorkBuddy/CodeBuddy offer does not establish current free access.
Can I self-host Tencent Hy4 Preview?
Tencent Hy4 Preview publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.
Model comparison catalogBenchmark library