Skip to main content

Compare Leaderboards Agents Insights

Loading...

Compare LLM models by quality, speed, and price. Find the right model for your use case.

Product

Explore Models
Compare
Leaderboards
LLM Selector
Cost Calculator
Provider Finder

Resources

Insights
FAQ
About
Methodology

Data Sources

Data from Artificial Analysis, OpenRouter, HuggingFace, and other public sources.

© 2026 FindLLM. All rights reserved.

Hy3-preview (Reasoning) — Tencent | FindLLM

Home
Explore
Hy3-preview (Reasoning)

Back to Explore

Hy3-preview (Reasoning)

Tencent·Released 2026-04-23

Related Models

Hy3-preview (Non-reasoning)2026-04-23 Tencent: Hy3 preview2026-04-22 Tencent: Hy3 preview (free)2026-04-22

Performance

85

tokens/sec

Faster than 49% of models

2.15

seconds

Faster than 20% of models

25.76

seconds

Faster than 19% of models

Market Median

86 tok/s

1% slower

Median TTFT

1.07s

101% slower

Speed Comparison

xAI: Grok 4.1 Fast

85 tok/s+1%

Reka Flash (Sep '24)

85 tok/s+1%

Hy3-preview (Non-reasoning)

84 tok/s-1%

Benchmarks

MMLU-ProNot evaluated

GPQA Diamond

86.7%

HLE

25.5%

LiveCodeBenchNot evaluated

SciCode

41.2%

TerminalBench Hard

34.1%

MATH-500Not evaluated

AIMENot evaluated

AIME 2025Not evaluated

IFBench

63.1%

Long Context Recall

54.7%

Tau2

92.7%

Market AverageTop Score

Quick Compare

Similar Models

MiniMax: MiniMax M2.5

MiniMax

Q: 41.9$0.40/1M197K ctx

GPT-5 (medium)

OpenAI

Q: 42.0$3.44/1M

Claude 4.1 Opus (Reasoning)

Anthropic

Q: 42.0$32.81/1M

DeepSeek V3.2 (Reasoning)

DeepSeek

Q: 41.7$0.34/1M

Qwen: Qwen3.5-27B

Alibaba

Q: 42.1$0.54/1M262K ctx

Z.ai: GLM 4.7

Z AI

Q: 42.1$0.74/1M203K ctx

Compare all 7 models

Quality Index

41.9

62nd of 507

Top 12%

Coding Index

36.5

69th of 417

Top 17%

Speed

85 tok/s

Top 51%

TTFT

2.15s

Market Position

Hy3-preview (Reasoning)Market Average