Hy4 flash
Tencent's low-cost Hy4 model for everyday office and agent work
Context window
1M
Input / 1M tokens
$0.07
Output / 1M tokens
$0.281
Provider
Tencent
Tencent Cloud TokenHub list price, per 1M tokens: $0.07 input, $0.281 output, $0.003 cached input (0.5, 2 and 0.02 yuan in the Guangzhou region). Inference is billed at 50% of list price until 12:00 on November 30, 2026 (Beijing time), and Hy4 flash is free in WorkBuddy until November 30. The pricing page does not list a maximum output length. · Data verified 2026-11-20
Hy4 flash (November 16, 2026) is the smaller, cheaper sibling of Tencent's flagship Hy4, aimed at everyday work such as research, document processing, slides, simple front-end pages and operating a PC; Tencent recommends the full Hy4 for long, complex software development. It takes multimodal input including images, has a 1M-token context and lists at $0.07/$0.281 per 1M tokens on Tencent Cloud, against $0.834/$2.501 for Hy4. Tencent says it slightly outperforms Hy4 Preview on public agent benchmarks and that its cost per task on its internal WorkIndex test is 1/20 of Hy4 and 1/8 of DeepSeek V4.1 Flash 1017, but it has not published the scores. Tencent plans to release the weights and has not given a date.
How to access it
Tencent Cloud TokenHub API (model ID hy4-flash) through the OpenAI Chat Completions, OpenAI Responses or Anthropic Messages protocols, and in Tencent products including WorkBuddy (free until November 30, 2026), Yuanbao, WeCom, QQ, ima and Marvis. Tencent says OpenRouter support is coming.
Strengths
- ✓$0.07/$0.281 per 1M tokens list price, half off on TokenHub until November 30, 2026
- ✓1M-token context with image input
- ✓Tencent reports half the output tokens per task after reinforcement learning, and decoding above 150 tokens per second
- ✓Works with the OpenAI Chat Completions, OpenAI Responses and Anthropic Messages protocols on TokenHub
Best for developers who...
When to choose it (and when not to)
Reach for Hy4 flash when...
- →You want a very cheap model with image input and a 1M context for office and agent tasks
- →You already use Tencent Cloud TokenHub or WorkBuddy
Look elsewhere if...
- ✕Long, complex software engineering (Tencent recommends the full Hy4)
- ✕You need published benchmark scores: Tencent has shared only internal cost comparisons
- ✕You need open weights today
How to use it
Quickstart
Pythonimport os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["TOKENHUB_API_KEY"],
base_url="https://tokenhub-intl.tencentcloudmaas.com/v1",
)
response = client.chat.completions.create(
model="hy4-flash",
messages=[{"role": "user", "content": "Summarize this report in five bullet points: ..."}],
)
print(response.choices[0].message.content)Tencent Cloud TokenHub is OpenAI-compatible (Chat Completions and Responses) and also accepts the Anthropic Messages protocol. Chinese mainland region: https://tokenhub.tencentcloudmaas.com/v1.
API model id: hy4-flash
Compare Hy4 flash
Compare Hy4 flash with any other model
Build a comparison →All model comparisons →