T
TENCENT / tencent/Hy4-flash-base

Tencent Hy4 Flash Base

Pretrained checkpoint in the Hy4 Flash family, published in the official repository on November 17. It retains the visual encoder and MoE architecture but is distinct from the instruction-tuned model and its FP8 format.

01 / SUMMARY

The essential

ORIENTATIONPretrained multimodal base model

Primary use stated or inferred cautiously from official documentation.

CONTEXT1,048,576 tokens in the weights configuration

Maximum input capacity when published by the source.

EXITText; generation maximum not specified for the base checkpoint

Documented maximum generation limit.

ACCESSHugging Face · downloadable checkpoint

Verified availability channels.

MODEL TYPEMultimodal · Text and reasoning

Primary functional family and verified specialties.

LICENSETencent Hy License

Declared terms for API access, model weights, or self-hosting.

02 / TECHNICAL SHEET

Limits and integration

API IDtencent/Hy4-flash-base
Model typeMultimodal · Text and reasoning
Access modelDownloadable weights
LicenseTencent Hy License
DeploymentSelf-hosted
Release17 NOV 2026
Knowledge cutoffNot published
EntranceText · Image
ExitText
Context window1,048,576 tokens in the weights configuration
maximum outputText; generation maximum not specified for the base checkpoint
ReasoningNot published
Published tools
Structured OutingsNot published
Batch processingNot published
Prompt cacheNot published
Fine-tuningNot published
Verified platformsHugging Face
BEST SUITED FOR

Research and adaptation of a base model with text and image input. The release includes weights and configuration. Pretraining measurements use few-shot protocols and must remain separate from assistant evaluations.

WORTH MONITORING

The assistant’s instructions, tools and results are not inherited. The Tencent Hy license has its own conditions and restrictions; it is not Apache-2.0. Deployment needs substantial infrastructure and a compatible runtime.

03 / CAPABILITIES

What it can do

01

Orientation

Research and adaptation of a base model with text and image input. The release includes weights and configuration. Pretraining measurements use few-shot protocols and must remain separate from assistant evaluations.

02

Context

1,048,576 tokens in the weights configuration · Text; generation maximum not specified for the base checkpoint

03

Tools and integration

Not published

04

Access

Hugging Face · downloadable checkpoint

Input modalities
TextImage
04 / PRICES

Documented cost

ConceptWorthUnit/condition
Standard inputNo published official API rateStandard API rate
Cached inputNot publishedReading reused prefixes
Cache write or storageNot publishedThe condition varies by provider
Standard outputSelf-hosted infrastructure costMay include reasoning tokens
Batch inputNot publishedAsynchronous processing
Batch outputNot publishedAsynchronous processing

No official dedicated API price has been verified. Downloading weights does not eliminate compute, memory or operating costs. TokenHub prices for hy4-flash apply to the assistant service, not this base checkpoint.

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

“

The pretraining table reports 71.1% on MMLU-Pro with five examples and macro-averaged accuracy. The figure is for Hy4 Flash Base, not the Instruct model or a test reproduced by Inferama.

BenchmarkResultMetricSource
MMLU-Pro (Hy4 Base 5-shot)71.1%Macro accuracy · 5-shotView 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-flash-base

Pretrained checkpoint published; revision 95b9f823 verified on November 18.

07 / SOURCES

Traceability