Tencent Releases Hy4 Flash, a Lower-Cost Multimodal Model With 1M-Token Context
This story is from 2026-11-16. It is preserved in the archive; the latest stories are on the live feed.
Tencent launched Hy4 Flash on November 16, 2026, a 220B-parameter multimodal MoE model with 12B active parameters, a one-million-token context window, and open weights including FP8 and base versions, priced substantially below full Hy4.
Read the full story at Unite.AI ↗
Timeline · 2 reports
- 2026-11-18 04:39 · AlphaSignal
Tencent Opens Hy4 Flash Weights With 1M-Token Context at 20x Lower Cost - 2026-11-16 10:25 · Unite.AI
Tencent Launches Hy4 Flash as a Lower-Cost Multimodal Model