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 mode…

tencent/hy3-preview-20260421

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 balance speed and depth depending on the task, while delivering strong code generation and reliable performance across multi-step, real-world workflows.

Specifications & capabilities

Input
text
Output
text
Context window · tokens
262144
Published model ID
Hy3-preview
Capabilities
promptCaching, reasoning, toolCalling
Description source ↗

Evaluations

41–60 / 86

LMArena

1,421.577

source id
lmarena
leaderboard name
LMArena Text Style Control Longer Query
dimension
style
metric
Arena rating
rank
141
sample size
3015
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1410.2781757252753
confidence upper
1432.8756520438524
Original result ↗

LMArena

1,387.609

source id
lmarena
leaderboard name
LMArena Text Factuality Industry Writing And Literature And Language
dimension
factuality
metric
Arena rating
rank
141
sample size
1247
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1371.7067061708397
confidence upper
1403.5115573605467
Original result ↗

LMArena

1,402.549

source id
lmarena
leaderboard name
LMArena Text Factuality Instruction Following
dimension
factuality
metric
Arena rating
rank
142
sample size
1972
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1389.6919757483465
confidence upper
1415.4050295208983
Original result ↗

LMArena

1,415.36

source id
lmarena
leaderboard name
LMArena Text Style Control Multi Turn
dimension
style
metric
Arena rating
rank
143
sample size
1202
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1397.814529629653
confidence upper
1432.9054046568262
Original result ↗

LMArena

1,413.04

source id
lmarena
leaderboard name
LMArena Text Style Control Industry Business And Management And Financial Operations
dimension
style
metric
Arena rating
rank
144
sample size
1419
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1397.102434924303
confidence upper
1428.9768737747877
Original result ↗

LMArena

1,400.495

source id
lmarena
leaderboard name
LMArena Text Spanish
dimension
general
metric
Arena rating
rank
145
sample size
207
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1360.6071181180669
confidence upper
1440.3832040481418
Original result ↗

LMArena

1,398.657

source id
lmarena
leaderboard name
LMArena Text Industry Business And Management And Financial Operations
dimension
general
metric
Arena rating
rank
145
sample size
1419
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1382.7928775149737
confidence upper
1414.5204521717467
Original result ↗

LMArena

1,432.738

source id
lmarena
leaderboard name
LMArena Text Style Control Hard Prompts
dimension
style
metric
Arena rating
rank
146
sample size
4539
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1423.6689885269182
confidence upper
1441.8060769857573
Original result ↗

LMArena

1,399.74

source id
lmarena
leaderboard name
LMArena Text Factuality Non English
dimension
factuality
metric
Arena rating
rank
146
sample size
3508
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1390.1611958924154
confidence upper
1409.3193372591713
Original result ↗

LMArena

1,430.649

source id
lmarena
leaderboard name
LMArena Text Style Control Industry Life And Physical And Social Science
dimension
style
metric
Arena rating
rank
147
sample size
1231
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1413.5129653923107
confidence upper
1447.7856606823857
Original result ↗

LMArena

1,435.224

source id
lmarena
leaderboard name
LMArena Text Chinese
dimension
general
metric
Arena rating
rank
148
sample size
453
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1406.4398631199267
confidence upper
1464.007879517024
Original result ↗

LMArena

1,402.158

source id
lmarena
leaderboard name
LMArena Text Longer Query
dimension
general
metric
Arena rating
rank
148
sample size
3015
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1390.7944284174134
confidence upper
1413.5216076772151
Original result ↗

LMArena

1,422.642

source id
lmarena
leaderboard name
LMArena Text Factuality English
dimension
factuality
metric
Arena rating
rank
149
sample size
3282
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1412.9638041184476
confidence upper
1432.3201345181822
Original result ↗

LMArena

1,412.272

source id
lmarena
leaderboard name
LMArena Text Hard Prompts
dimension
general
metric
Arena rating
rank
150
sample size
4539
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1403.1842507314423
confidence upper
1421.3596411991437
Original result ↗

LMArena

1,397.209

source id
lmarena
leaderboard name
LMArena Text Style Control Non English
dimension
style
metric
Arena rating
rank
150
sample size
3509
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1387.0130539820163
confidence upper
1407.4050777752316
Original result ↗

LMArena

1,433.904

source id
lmarena
leaderboard name
LMArena Text Factuality Hard Prompts
dimension
factuality
metric
Arena rating
rank
151
sample size
4538
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1425.3858303962115
confidence upper
1442.4216152978547
Original result ↗

LMArena

1,390.029

source id
lmarena
leaderboard name
LMArena Text Non English
dimension
general
metric
Arena rating
rank
151
sample size
3509
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1379.8559901300055
confidence upper
1400.2024253338664
Original result ↗

LMArena

1,422.723

source id
lmarena
leaderboard name
LMArena Text Industry Software And It Services
dimension
general
metric
Arena rating
rank
152
sample size
2775
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1411.285921663978
confidence upper
1434.1595509366268
Original result ↗

LMArena

1,420.925

source id
lmarena
leaderboard name
LMArena Text Coding
dimension
general
metric
Arena rating
rank
152
sample size
2035
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1407.4802078196976
confidence upper
1434.370575798676
Original result ↗

LMArena

1,412.202

source id
lmarena
leaderboard name
LMArena Text Factuality Overall
dimension
factuality
metric
Arena rating
rank
152
sample size
6813
evaluated at
2026-11-09T00:00:00.000Z
source model id
Hy3 preview
confidence lower
1405.2813614997463
confidence upper
1419.123229356545
Original result ↗

Published prices

1–3 / 3

tencent/hy3-preview ↗

$0.06

Inference provider
openRouter
Platform
OpenRouter
Unit
cache_read_tokens_per_million
Currency
USD
Billing conditions

tencent/hy3-preview ↗

$0.18

Inference provider
openRouter
Platform
OpenRouter
Unit
input_tokens_per_million
Currency
USD
Billing conditions

tencent/hy3-preview ↗

$0.60

Inference provider
openRouter
Platform
OpenRouter
Unit
output_tokens_per_million
Currency
USD
Billing conditions

Provider specifications & access

1–1 / 1

Hy3-preview ↗

source model id
tencent/Hy3-preview
hosting
siliconflow-hosted
platforms
  • SiliconFlow
architecture
input_modalities
  • text
output_modalities
  • text
capabilities
  • promptCaching
  • toolCalling

Source specification observations

1–3 / 3

openRouter ↗

field
context_length
value
262144
normalized value
262144
scope
source-model
source id
openRouter
confidence
1

siliconFlow ↗

field
context_length
value
262K
scope
source-model
source id
siliconFlow
confidence
1

openRouter ↗

field
max_output_tokens
value
235929
normalized value
235929
scope
source-model
source id
openRouter
confidence
1

Original sources

1–12 / 12

SiliconFlow · Model page ↗

hy.tencent.com · Model reference ↗

Hugging Face · Model repository ↗

OpenRouter · Model page ↗

SiliconFlow · Playground ↗

SiliconFlow · API documentation ↗

Epoch AI · Catalog ↗

Hugging Face · Catalog ↗

Hugging Face · Catalog ↗

LMArena · Catalog ↗

OpenRouter · Catalog ↗

SiliconFlow · Catalog ↗

Source links

1–8 / 8

Explore more models

Models

Loading
EVIDENCE / COMPARISON

Benchmarks

Explore every benchmark

Loading

These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

Install TensorTrace