Key Notes
- Tencent released Hy4 Flash on November 16 with text and image input, approximately one million tokens of context and tools for agent and office workflows.
- Standard API prices are $0.07 for input and $0.281 for output per million tokens, with a 50% launch discount through November 30.
- Tencent’s lower task-cost and model-quality claims come from internal evaluations, while OpenRouter support and release of the model’s weights remain announced plans.
Tencent released Hy4 Flash on November 16, adding a lower-cost model for AI agents, document work and everyday coding assistance. The model accepts text and images, produces text, and supports a context window of roughly one million tokens.
The company introduced the model as a smaller companion to Hy4. Hy4 Flash has 224 billion total parameters, with 12 billion active per token.
Low API rates and a launch discount
Tencent Cloud’s English-language model page lists standard rates of $0.07 per million input tokens and $0.281 per million output tokens. Cached input is listed at $0.003 per million tokens.
A 50% launch promotion reduces input and output rates to $0.035 and $0.1405 respectively. Tencent says the API discount runs through November 30; Hy4 Flash is also free to use in WorkBuddy during the launch period.
The model supports tool calling, structured outputs and prompt caching. TokenHub lists compatibility with Chat Completions, Responses and Messages API formats, giving developers several ways to connect it to existing tools.
Tencent’s task-cost claims come from internal tests
On WorkIndex, Tencent’s internal test of office and agent workflows, the company reports task costs around one-twentieth of Hy4’s and performance slightly above Hy4 Preview.
Tencent also reports satisfaction close to Anthropic’s Claude Haiku 5.5 in blind tests involving 85 internal users and 259 tasks. Those findings describe its own employee sample and workflows.
Cost per finished task depends on more than the advertised token rate. Response length, tool calls and repeated attempts also affect the bill. Tencent’s task-level results are useful evidence about its chosen workflows, while developers evaluating a switch will need results from their own work.
Available now, with open weights planned
Hy4 Flash is available through TokenHub, WorkBuddy and Yuanbao. Tencent says OpenRouter support and open weights are coming; the announcement does not establish that those releases have happened.
Tencent recommends Flash for common office jobs and simple frontend demos, while continuing to recommend Hy4 for demanding, long software-engineering tasks.
The release adds another option to a busy week for Chinese models, following Qwen4-Max. Hy4 Flash’s immediate appeal is its combination of multimodal input, long context and low API rates for workflows that call a model repeatedly.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.