Tier
Major
Headquarters
Shenzhen, China
Founded
1998
Flagship
Hunyuan
Models tracked
2
Artificial Analysis
Top intelligence
41.9
Artificial Analysis Index
Cheapest model
$0.20
blended /1M
Fastest output
99
tokens / sec
Flagship
Tencent’s top-scoring model on the Artificial Analysis Intelligence Index.
Most intelligent model
Hy3-preview (Reasoning)
Released 2026-04-23
Intelligence
41.9
Coding
36.5
Math
—
In $/1M
$0.12
Out $/1M
$0.43
Speed t/s
99
Benchmark breakdown
Models & benchmarks
Every Tencent model on file — intelligence, coding, price, and speed.
| Model | Released | Intel | Coding | Math | In $/1M | Out $/1M | Speed t/s | Latency s |
|---|---|---|---|---|---|---|---|---|
| Hy3-preview (Reasoning) | 2026-04-23 | 41.9 | 36.5 | — | $0.12 | $0.43 | 99 | 2.53 |
| Hy3-preview (Non-reasoning) | 2026-04-23 | 33.7 | 34.3 | — | $0.12 | $0.43 | 99 | 2.61 |
Trajectory
Tencent’s model releases over time, by intelligence.
House strengths
Where Tencent’s lineup over-indexes, averaged across its models.
Capability profile
Best value
Hy3-preview (Reasoning)
Most intelligence per dollar — 41.9 index at $0.20/1M blended.
209 index points per $1
Media models
Image, video, and speech models, ranked by arena Elo.
| Model | Type | Released | Elo | Rank |
|---|---|---|---|---|
| HunyuanImage 3.0 Instruct (Fal) | image editing | 2026-01 | 1233 | #7 |
| HunyuanImage 3.0 (Fal) | text to image | 2025-09 | 1124 | #49 |
| HunyuanVideo-1.5 (Fal) | image to video | 2025-11 | 1122 | #62 |
| HunyuanImage 3.0 Instruct (Fal) | text to image | 2026-01 | 1112 | #56 |
| SRPO | text to image | 2025-09 | 1074 | #79 |
| HunyuanImage 2.1 | text to image | 2025-09 | 1068 | #85 |
| HunyuanVideo-1.5 (Fal) | text to video | 2025-11 | 1029 | #64 |
| Hunyuan Video (Fal) | text to video | 2024-12 | 1002 | #70 |
| Hunyuan Video (Fal) | image to video | 2024-12 | 909 | #77 |
Overview
The technology and gaming conglomerate behind the Hunyuan family of text, image, and video generation models.
Visit Tencent