Tencent: Hy3 preview

Tencent: Hy3 preview

tencent · Released Apr 22, 2026 Professional
Intelligence #10 / 576
82.0 Our Score
Speed #127 / 271
98.9 tokens / sec
Input #171 / 577
$0.063 per 1M tokens
Output #197 / 577
$0.210 per 1M tokens
Context #108 / 577
262,144 tokens

Analysis Summary

Tencent Hy3 preview is a cost-efficient model with an intelligence index of 33.7 and a strong agentic index of 49.7, supported by tool use and function calling across a 262K token context window. GPQA performance at 0.732 and a tau2 score of 0.675 indicate reasonable real-world task reliability, though coding capability is moderate.

For businesses, the model's agentic strengths make it a candidate for multi-step automation and tool-orchestration workflows where cost is a priority. At $0.063 input and $0.21 output per million tokens, it offers strong value for the capability level. Long-context reasoning (LCR 0.337) is adequate but not a standout.

Hy3 preview suits teams running high-volume agentic or structured-output tasks on a tight budget. For reasoning-heavy or coding-intensive work, stronger models in the 40+ intelligence range would be more appropriate.

Assessed June 6, 2026

Editorial notes

Tencent Hy3 preview delivers strong agentic performance and good reasoning at very low cost, with tool use and a 262K context window, though coding capability is moderate.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence4.6Technical5.8Value8.3Content4.6
Intelligence 4.6/10
Technical 5.8/10
Content 4.6/10
Value 8.3/10

How Tencent: Hy3 preview compares

Tencent: Hy3 preview ranks #49 of 378 AI models we track for overall intelligence, #55 of 315 for coding, #43 of 289 for agentic tasks. Its 262K-token context window is larger than 81% of the models we list. At $0.06 per million input tokens it is cheaper than 70% of comparable models.

About Tencent: Hy3 preview

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..

Capabilities

Tool Use Function Calling

Performance Indices

Source: Artificial Analysis

41.9 Intelligence Index
36.5 Coding Index
63.4 Agentic Index

Benchmark Scores

Intelligence

GPQA Diamond 86.7% Graduate-level scientific reasoning
HLE 25.5% Humanity's Last Exam
SciCode 41.2% Scientific computing

Technical

TerminalBench Hard 34.1% Agentic terminal tasks
τ²-Bench 92.7% Conversational agent benchmark

Content

IFBench 63.1% Instruction following
LCR 54.7% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does Tencent: Hy3 preview stack up?

Compare side-by-side with other professional models.

Compare Models

Model Information

OpenRouter ID tencent/hy3-preview
Providertencent
Release Date April 22, 2026
Context Length262,144 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.06 $0.000063
Output $0.21 $0.000210

Live Performance

Live endpoint metrics, refreshed every 30 minutes.

98.3%
Avg Uptime
4,499ms
Best Latency (TTFT)
44 tok/s
Best Throughput
2/2
Active Endpoints
Available via: GMICloud, SiliconFlow

Leaderboard Categories

Frequently asked questions about Tencent: Hy3 preview

How much does Tencent: Hy3 preview cost?

Tencent: Hy3 preview costs $0.06 per million input tokens and $0.21 per million output tokens.

What is the context window of Tencent: Hy3 preview?

Tencent: Hy3 preview has a context window of 262,144 tokens (262K).

Is Tencent: Hy3 preview good for coding?

On our coding benchmark index, Tencent: Hy3 preview ranks #55 of 315 models, placing it in the top quartile of the field for code generation and debugging.

What can Tencent: Hy3 preview do?

Tencent: Hy3 preview supports tool use and function calling.

Who created Tencent: Hy3 preview?

Tencent: Hy3 preview is developed by Tencent and was released on April 22, 2026.