Tencent FRONTIER API · Context: 1,048,576 tokens (1M)

Tencent Hy4 flash

Tencent's low-cost Hy model for everyday office, research and light computer-use work: ~$0.07 / $0.28 per 1M tokens, >150 tok/s, and a WorkIndex score slightly above the 770B Hy4 Preview at about 1/13 of its per-task cost. Weak on hard coding (19.2 Terminal-Bench 4.0, vendor-reported).

Provider
Tencent (API now; open weights announced (not released, no license named as of Nov 16, 2026))
Architecture
Hybrid attention + LatentMoE + Engram lookup module, multimodal input; 224B total (incl. 64B Engram) / 12B active; KV cache 1/6 of Hy4
Parameters
224B Total (64B Engram) / 12B Active
Context Window
1,048,576 tokens (1M) (Max Output: 0 tokens)
API Pricing
Input: $0.07 / 1M · Cached: $0.00 / 1M · Output: $0.28 / 1M
Hardware Target
API only until weights ship

Interactive Token Cost Estimator (Hy4 flash)

Est. API Cost (0% Cache): --/mo | With 70% Prompt Cache: --/mo
Compare all 12 models →

Architectural Strengths & Target Workloads

  • Very low price with cached input at ~$0.003 / 1M
  • >150 tok/s decode and 50% fewer output tokens per task after RL
  • WorkIndex 42.1 vs 39.7 for Hy4 Preview; blind-test satisfaction close to Claude Haiku 5.5

Primary Engineering Workloads: Cheap Executor Model for Agent Pipelines · Docs, Slides & Spreadsheet Agents · Lightweight Computer Use.

Frequently Asked Questions (Hy4 flash)

How much does Tencent Hy4 flash cost per 1M tokens?

Tencent Hy4 flash is priced at $0.07 per 1M uncached input tokens, $0.00 per 1M cached input tokens, and $0.28 per 1M output tokens (Tencent Cloud TokenHub (model id hy4-flash; CNY 0.50 / 2.00 / 0.02 list, 50% off until Nov 30, 2026), WorkBuddy (free until Nov 30), Yuanbao, WeCom, QQ, ima, Marvis; OpenRouter "coming soon").

What is the context window of Tencent Hy4 flash?

It supports up to 1,048,576 tokens (1M) (1,048,576 tokens) and a maximum generation output of 0 tokens.

Full TechPulse Engineering Analysis

Tencent Hy4 flash: 12B Active Parameters, 1M Context, and $0.07 per Million Input Tokens

Tencent's new small Hy model costs about a twentieth of Hy4 per task on Tencent's office-work benchmark and decodes at more than 150 tokens per second. Tencent says to use it for documents, slides, research and light computer use, and to keep Hy4 for serious software engineering.

Read Full Technical Dispatch (4 min read) →
Risograph-style illustration of a small terracotta cube moving along a long thin rail with soft speed lines behind it, three small ochre coins beside it, and a tall slate tower standing in the distance