modelgrep
T

Tencent: Hy3 preview

tencent/hy3-preview

36th smartest of 180Cheaper than 91% of paidReasoningTools
Use via OpenRouter ↗
Intelligence
41.9
36th of 180
Design Elo
Speed
62
138th fastest
Latency
3.8s
first token
Input price
$0.063
28th cheapest
Context
262K

How it compares

Smarter than80%
of all ranked models
Faster than55%
of all ranked models
Cheaper than91%
of all ranked models

Overview

Hy3 preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels across disabled, low, and high modes, allowing it to...

Benchmarks

independent · via OpenRouter
Artificial Analysis77th percentile
Intelligence Index
41.9
Coding Index
36.5
Agentic Index
55.7
GPQA Diamond
87%
Humanity's Last Exam
26%
SciCode
41%
Tau²-Bench (agentic)
93%

Providers & pricing (2)

ProviderIn $/MOut $/MUptime
GMICloudbf16$0.063$0.21075.5%
SiliconFlow$0.066$0.26095.8%

Specifications

Context window262K
Max output
Knowledge cutoff
Input modalitiestext
Output modalitiestext
Prompt caching
Cache read price$0.021/M
ModeratedNo

Hy3 preview FAQ

How much does Hy3 preview cost?

Hy3 preview costs $0.063 per million input tokens and $0.210 per million output tokens via OpenRouter, making it 28th cheapest of 298 paid models.

How smart is Hy3 preview?

Hy3 preview scores 41.9 on the Artificial Analysis Intelligence Index, ranking 36th of 180 benchmarked models, with a GPQA Diamond score of 87%.

How fast is Hy3 preview?

Hy3 preview generates around 62 tokens per second with 3.8s time-to-first-token (p50), the 138th fastest tracked model.

What is Hy3 preview's context window?

Hy3 preview supports a 262K-token context window. It accepts text input.

Compare head-to-head