NOUMENOpen app
NOUMEN

tencent · Open weights

Tencent Hy4 Preview

Added Aug 28, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Hy4 Preview is Tencent's 770B-parameter mixture-of-experts model with 49B active parameters. It is designed for coding agents, complex tool-use workflows, and productivity tasks, with a 1M-token context window and configurable reasoning effort.

  • Reasoning
  • Tools
  • Structured output
Input
Text
Output
Text
Reasoning effort
none · low · high
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.83
Output
$2.50
Cache read
$0.042

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
64K
Average output (7 days)
2K tokens
Longer than 92% of 468 models

Side by side

Tencent Hy4 Preview compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Tencent Hy4 PreviewtencentThis modelTencent Hy3tencentRemove
Providertencenttencent
Discovery tierOpen weightsEfficient
AddedAug 28, 2026Jul 6, 2026
Limits
Context window1.05MBest262.14K
Max output64K128KBest
Average output2K tokens1.4K tokens
Pricing, USD per 1M tokens
Input$0.83$0.066Best
Output$2.50$0.26Best
Cache read$0.042$0.029Best
Performance
Throughput40.8 tok/s61.4 tok/sBest
Time to first token4.3 s3.5 sBest
Request success (24h)99.8%Best99.5%
Benchmark indexes
Intelligence Index—25.3
Coding Index—58.8
Capabilities
ReasoningYesYes
ToolsYesYes
Structured outputYesYes
Reasoning effortnone, low, highnone, low, high

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Tencent Hy4 Preview
  • Tencent Hy3

Intelligence Index

  • Tencent Hy4 Preview—
  • Tencent Hy325.3
3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 58.8

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 18.4%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 2.7%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 947 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 1,063 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 11.2%

  • AA-LCRv1.1

    Reasoning across long documents

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 79.0%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 33.5%

  • GPQA Diamond

    Graduate-level science questions

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 89.7%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 0.5%

  • SciCode

    Scientific computing in Python

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 48.6%

Legacy evaluations

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Tencent Hy4 Preview: —
    • Tencent Hy3: 79.0%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
40.8 tok/s
Faster than 30% of 259 models
Time to first token
4.3 s
Quicker than 40% of 259 models
Prompt cache hits
35.1%
About $0.56 per 1M input tokens with caching
Request success (24h)
99.8%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 88.8%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tencent/hy4-preview",
    "reasoning_effort": "none",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="tencent/hy4-preview",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

tencent

Other tencent models, starting with the same family.

  • EfficientCompared
    Tencent Hy3

    262.1K context$0.066 / $0.26