Inference
EN
Tencent

Hy3

Hy3 is a reasoning model from Tencent. It has 7 scored results in 4 categories. It has no index.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-07
UnrankedEnough results for an index.
IndexAll categories, out of 100.UnrankedMedian 51.0
CoverageIndex weight that has results.70%
SpeedOutput tokens per second.49/sMedian 67/s
First tokenSeconds to the first output token.3.0 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.132Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.528Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.1439±6 #79Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
tencent/hy3
Max outputMaximum output tokens per request.
128K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.033
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
4 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
OpenRouterRouter$0.132$0.528262K
SiliconFlowHost$0.132$0.528262K
Deep InfraDeprecatedHost$0.13$0.53262K
Novita AIHost$0.14$0.58262K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted57.4—
CodingCode writing and repair.1 counted59.2—
ReasoningLogic problems and puzzles.2 counted53.9—
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.2 counted58.7—
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.None counted——

Results

7 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis GPQA DiamondKnowledge89.7%39 / 199
Artificial Analysis Long Context ReasoningReasoning79.0%58 / 217
Artificial Analysis Humanity's Last ExamKnowledge33.5%68 / 222
Critical Physics TasksReasoning4.9%84 / 214
Artificial Analysis SciCodeCoding48.6%55 / 114
GDPval-AA normalizedAgentic28.2%66 / 133
Agent Arena steerabilityAgentic−3.138 / 54
Agent Arena command recoveryAgentic−7.244 / 50
Agent Arena task outcomeAgentic−11.848 / 54

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseLMArenaCC BY 4.0models.devMIT

History

5 releases
Output / 1M$0.528 → $0.528
23 Oct 202612 Nov 2026

Questions

How much does Hy3 cost?
$0.132 per 1M input tokens, $0.528 per 1M output tokens.
How fast is Hy3?
49 output tokens per second, 3.0 s to the first token.
What is the context window of Hy3?
262K tokens.
Is Hy3 open source?
Yes. Its weights are published.
When was Hy3 released?
July 2026.

More from Tencent

All models
Hy ASR 3.0 previewUnrankedUI-Mate-27BUnrankedUI-Mate-9BUnrankedAuKUnrankedHy MT2 7BUnranked