Skip to content
llmwaves

Hy3 preview

Tencent · released Apr 22, 2026

Open weightsReasoningTool use

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...

Specification

Context window
262K
Max output
236K
Knowledge cutoff
Not stated
Parameters
298.8B
Licence
other
Serving providers
1
Moderated
No
Uptime
100.0%

Intelligence

22.7

71th percentile

Coding

—

Coding Index

Agentic

24.1

Agentic Index

Output speed

81 t/s

Median across providers

Latency

2.29s

Time to first token

Cost per task

$0.06

Estimated

Benchmarks

Where the score comes from

The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.

Evaluation scores

Percentage correct · higher is better

  • τ²-bench (Telecom)
    92.7%
  • GPQA Diamond
    86.7%
  • AA-LCR (long context)
    64.7%
  • IFBench
    63.1%
  • Terminal-Bench Hard
    34.1%
  • Humanity's Last Exam
    27.8%

An evaluation missing from this list was not run for this model — it is not a zero.

View as table
EvaluationScore
τ²-bench (Telecom)92.7%
GPQA Diamond86.7%
AA-LCR (long context)64.7%
IFBench63.1%
Terminal-Bench Hard34.1%
Humanity's Last Exam27.8%

Against its peers

Intelligence Index · this model highlighted, nearest peers in grey

Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.

View as table
ModelIntelligence
GPT-523.0
Qwen3.5-27B22.9
M2.722.8
M2.522.8
Hy3 preview22.7
Gemini 3.5 Flash Lite22.2
GLM 4.722.2
Kimi K2 Thinking22.0

Percentile among all indexed models

Intelligence71th
Agentic57th
Terminal work78th

Pricing

What it costs to run

List prices per million tokens, plus what one representative task works out to.

List price

Input / 1M tokens
$0.18
Output / 1M tokens
$0.6
Cached input / 1M
$0.06
Blended 3:1
$0.285

One task, estimated

$0.06

Input tokens
50,000
Output tokens
80,000
Profile
Reasoning

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).

Serving providers

Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.