llmboard.aiAI model intelligence
Home

Model Rankings

OverallOpen ModelsAgentCodingReasoningMathKnowledgeInstruction FollowingTextVision
Image GenerationImage Editing
Video GenerationImage to VideoVideo Editing
Text to SpeechSpeech to Text
Embeddings

Efficiency

Chat Token PricingImage PricingVideo PricingAudio Pricing
Chat Speed & LatencyProvider Reliability

Benchmarks

GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-Pro
All Benchmarks

Tools

Model Directory

Scoring & Data

Scoring & Data
1224 models729 benchmarks

Leaderboard Center

Overall RankingCodingCore BenchmarksPrice & ValueRuntime Performance

Modalities

All ModelsImage GenerationImage EditingVideo GenerationImage-to-VideoVideo EditingText-to-SpeechSpeech-to-TextEmbeddings

Data & Methods

Scoring MethodAll Benchmarks
llmboard.aiCopyright 2026 llmboard.ai

Tencent model product

Hy3

Hy3 is a 295B-parameter Mixture-of-Experts (MoE) model from Tencent with 21B active parameters and a 256K context window.

Updated Sep 8, 2026. Default version: Hy3

LLMBoard Score71.3Hy3
Coverage20%27 benchmark families
Context windowN/ATokens
Official input priceN/AOfficial price unavailable

On this page

  • Capability
  • Benchmarks
  • Arena
  • Pricing
  • Runtime
  • Specification
  • Versions
  • Similar models
  • About
  • FAQ

Hy3 Capability Profile

This profile uses the model's current scored version. Arena ratings and prices are shown separately.

Hy3 LLMBoard score breakdown

Hy3 Benchmark Results

Benchmark scores for Hy3.

30 of 41 rows
Columns

Show columns

Sort by
Benchmark
Score
Rank
Participants
Percentile
Evidence
Evaluated
BenchmarkCL-benchScore23.80%Rank01Participants2Percentile100.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkCL-bench (Life)Score17.00%Rank01Participants1Percentile100.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkCMT-BenchmarkScore37.90%Rank01Participants1Percentile100.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkHumanity's Last Exam (no tools, text-only)Score47.00%Rank01Participants3Percentile100.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkPHYBenchScore77.40%Rank01Participants1Percentile100.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkAA-LCRScore73.40%Rank02Participants19Percentile94.44%EvidenceCEvaluatedSep 8, 2026
BenchmarkHorizonMathScore7.10%Rank02Participants4Percentile66.67%EvidenceCEvaluatedSep 8, 2026
BenchmarkArXivMathScore52.20%Rank03Participants3Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkFrontierScience OlympiadScore74.80%Rank03Participants3Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkIMO-AnswerBenchScore90.00%Rank03Participants20Percentile89.47%EvidenceCEvaluatedSep 8, 2026
BenchmarkUSAMO 2026Score30.24 pointsRank03Participants3Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkWildClawBenchScore53.60%Rank03Participants5Percentile50.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkClaw-EvalScore68.50%Rank04Participants14Percentile76.92%EvidenceCEvaluatedSep 8, 2026
BenchmarkDeepSearchQAScore91.00%Rank04Participants10Percentile66.67%EvidenceCEvaluatedSep 8, 2026
BenchmarkFrontierScience ResearchScore21.30%Rank04Participants4Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkSkillsBenchScore55.30%Rank04Participants9Percentile62.50%EvidenceCEvaluatedSep 8, 2026
BenchmarkSuperChemScore54.90%Rank04Participants4Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkHumanity's Last Exam (with tools, text-only)Score53.20%Rank05Participants5Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkWideSearchScore76.40%Rank05Participants11Percentile60.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkMathArena ApexScore38.70%Rank06Participants9Percentile37.50%EvidenceCEvaluatedSep 8, 2026
BenchmarkAPEX-AgentsScore25.60%Rank09Participants9Percentile0.00%EvidenceCEvaluatedSep 8, 2026
BenchmarkMCP AtlasScore79.10%Rank09Participants34Percentile75.76%EvidenceCEvaluatedSep 8, 2026
BenchmarkDeepSWEScore28.00%Rank12Participants13Percentile8.33%EvidenceCEvaluatedSep 8, 2026
BenchmarkNL2RepoScore45.60%Rank13Participants22Percentile42.86%EvidenceCEvaluatedSep 8, 2026
BenchmarkSWE-bench MultilingualScore75.80%Rank13Participants43Percentile71.43%EvidenceCEvaluatedSep 8, 2026
BenchmarkBrowseCompScore84.20%Rank15Participants63Percentile77.42%EvidenceCEvaluatedSep 8, 2026
BenchmarkSWE-Bench VerifiedScore78.00%Rank22Participants113Percentile81.25%EvidenceCEvaluatedSep 8, 2026
BenchmarkTerminal-Bench 2.1Score71.70%Rank22Participants35Percentile38.24%EvidenceCEvaluatedSep 8, 2026
BenchmarkSWE-Bench ProScore57.90%Rank24Participants55Percentile57.41%EvidenceCEvaluatedSep 8, 2026
BenchmarkGPQAScore90.40%Rank25Participants247Percentile90.24%EvidenceCEvaluatedSep 8, 2026

Hy3 Arena Results

Preference and agent-evaluation results for the default version.

30 of 78 rows
Columns

Show columns

Sort by
Arena
Category
Rank
Rating / score
Votes
Observations
Result date
ArenatextCategoryindustry mathematicalRank13Rating / score1,493.94Votes387ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryindustry mathematicalRank13Rating / score1,504.54Votes387ObservationsN/AResult dateSep 2, 2026
Arenatext factualityCategoryindustry life and physical and social scienceRank19Rating / score1,491.86Votes1,035ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryindustry life and physical and social scienceRank19Rating / score1,493.88Votes1,050ObservationsN/AResult dateSep 2, 2026
Arenatext factualityCategoryrussianRank23Rating / score1,467.54Votes709ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryindustry medicine and healthcareRank23Rating / score1,488.16Votes459ObservationsN/AResult dateSep 2, 2026
Arenaagent steerabilityCategoryoverallRank25Rating / score0.00VotesN/AObservations11.8KResult dateSep 5, 2026
ArenatextCategorymathRank26Rating / score1,468.12Votes359ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategorychineseRank26Rating / score1,516.58Votes465ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategorymathRank28Rating / score1,475.24Votes359ObservationsN/AResult dateSep 2, 2026
ArenawebdevCategorywebdev-reactRank29Rating / score1,517.75Votes2,224ObservationsN/AResult dateSep 5, 2026
Arenaagent bash recovery stepsCategoryoverallRank32Rating / score-0.01VotesN/AObservations22.6KResult dateSep 5, 2026
Arenatext style controlCategoryspanishRank32Rating / score1,461.26Votes268ObservationsN/AResult dateSep 2, 2026
Arenaagent praise complaintCategoryoverallRank33Rating / score-0.04VotesN/AObservations3.7KResult dateSep 5, 2026
ArenatextCategoryindustry life and physical and social scienceRank33Rating / score1,466.80Votes1,050ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryfrenchRank33Rating / score1,484.93Votes302ObservationsN/AResult dateSep 2, 2026
ArenawebdevCategoryoverallRank33Rating / score1,511.65Votes3,173ObservationsN/AResult dateSep 5, 2026
ArenawebdevCategorywebdevRank33Rating / score1,511.65Votes3,173ObservationsN/AResult dateSep 5, 2026
ArenaagentCategoryoverallRank34Rating / score-0.02VotesN/AObservations876.8KResult dateSep 5, 2026
ArenatextCategorychineseRank35Rating / score1,499.75Votes465ObservationsN/AResult dateSep 2, 2026
ArenatextCategoryrussianRank36Rating / score1,451.41Votes745ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryexpertRank36Rating / score1,489.67Votes709ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategoryrussianRank36Rating / score1,465.70Votes745ObservationsN/AResult dateSep 2, 2026
ArenatextCategoryindustry medicine and healthcareRank37Rating / score1,459.36Votes459ObservationsN/AResult dateSep 2, 2026
ArenatextCategoryfrenchRank39Rating / score1,463.39Votes302ObservationsN/AResult dateSep 2, 2026
ArenatextCategoryspanishRank39Rating / score1,446.96Votes268ObservationsN/AResult dateSep 2, 2026
Arenaagent task outcome explicitCategoryoverallRank40Rating / score-0.06VotesN/AObservations11.7KResult dateSep 5, 2026
ArenawebdevCategorywebdev-htmlRank41Rating / score1,476.61Votes350ObservationsN/AResult dateSep 5, 2026
Arenatext style controlCategoryenglishRank42Rating / score1,467.69Votes2,697ObservationsN/AResult dateSep 2, 2026
Arenatext style controlCategorylonger queryRank43Rating / score1,466.00Votes3,165ObservationsN/AResult dateSep 2, 2026

Hy3 Pricing

Official vendor API pricing appears first, followed by individual provider offers.

Official API
N/A
Official provider
N/A
Lowest third-party
From $0.066 input, $0.26 output per 1M via NanoGPT
Tracked offerings
15
15 rows
Columns

Show columns

Sort by
Provider
Provider model ID
Region
Input / 1M
Output / 1M
Context
Updated
ProviderTencent TokenHubProvider model IDhy3RegionglobalInput / 1MN/AOutput / 1MN/AContext256KUpdatedSep 8, 2026
ProviderKenariProvider model IDhy3RegionglobalInput / 1MN/AOutput / 1MN/AContext256KUpdatedSep 8, 2026
ProviderTencent Token PlanProvider model IDhy3RegionglobalInput / 1MN/AOutput / 1MN/AContext256KUpdatedSep 8, 2026
ProviderNanoGPTProvider model IDtencent/hy3RegionglobalInput / 1M$0.066Output / 1M$0.26Context262.1KUpdatedSep 8, 2026
ProviderOpenRouterProvider model IDtencent/hy3RegionglobalInput / 1M$0.132Output / 1M$0.528Context262.1KUpdatedSep 8, 2026
ProviderKilo GatewayProvider model IDtencent/hy3RegionglobalInput / 1M$0.132Output / 1M$0.528Context262.1KUpdatedSep 8, 2026
ProviderJalapeno CloudProvider model IDHy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context202.8KUpdatedSep 8, 2026
ProviderDeep InfraProvider model IDtencent/Hy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context262.1KUpdatedSep 8, 2026
ProviderHugging FaceProvider model IDtencent/Hy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context262.1KUpdatedSep 8, 2026
ProviderOpenCode GoProvider model IDhy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context256KUpdatedSep 8, 2026
ProviderRequestyProvider model IDhy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context262.1KUpdatedSep 8, 2026
ProviderVercel AI GatewayProvider model IDtencent/hy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context262.1KUpdatedSep 8, 2026
ProviderDevPass (LLM Gateway)Provider model IDhy3RegionglobalInput / 1M$0.14Output / 1M$0.58Context262.1KUpdatedSep 8, 2026
ProviderCrossModelProvider model IDtencent/hy3RegionglobalInput / 1M$0.16Output / 1M$0.64Context262.1KUpdatedSep 8, 2026
ProviderOrcaRouterProvider model IDtencent/hy3RegionglobalInput / 1M$0.18Output / 1M$0.59Context256KUpdatedSep 8, 2026

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

Hy3 Runtime Performance

Provider-specific output speed and catalog latency for Hy3. Runtime does not affect the capability score.

No runtime data

No provider-specific speed or latency record is linked to the default version yet.

Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.

Hy3 Specifications

Technical details for the model's default version.

Version
Hy3
Released
Jul 6, 2026
Knowledge cutoff
Unknown
Parameters
295B
Context window
N/A
Max output
N/A
Inputs
text
Outputs
text
Open weights
Yes
License
Apache 2.0

Hy3 Versions

Available versions of this model. The score column identifies the version used in the overall ranking.

1 row
Columns

Show columns

Sort by
Version
Released
LLMBoard
Parameters
Context
Max output
Open weights
License
VersionHy3ReleasedJul 6, 2026LLMBoard71.35Parameters295BContextN/AMax outputN/AOpen weightsYesLicenseApache 2.0

Models similar to Hy3

Recommendations prioritize the same model type and family, then the closest LLMBoard score.

#16+12.36
TE

Hy4

Tencent

83.71 LLMBoard

Details
#43+0.02
AN

Claude Sonnet 4.6

Anthropic

71.37 LLMBoard

Details
#42+0.15
AC

Qwen3.7 Plus

Alibaba Cloud / Qwen Team

71.50 LLMBoard

Details
#45-0.32
OP

GPT-5.2-Pro

OpenAI

71.03 LLMBoard

Details
#46-0.41
SA

Sakana Namazu

Sakana AI

70.94 LLMBoard

Details
#40+0.74
ME

Muse Spark

Meta

72.09 LLMBoard

Details

What is Hy3?

Key information about Hy3 and its available data.

Hy3 is a hybrid-thinking large language model developed by the Tencent Hy Team, with configurable reasoning effort modes: no_think, low, and high. 8B MTP layer.

Data as of 2026-09-08.

FAQ

Common questions about Hy3.

When was Hy3 released?

Hy3's default version was released on Jul 6, 2026.

How much does Hy3 cost?

No official standard PAYG price is currently available for Hy3. The lowest tracked third-party offer starts at $0.066 input and $0.26 output via NanoGPT.

Who created Hy3?

Hy3 was created by Tencent.

What is the context window for Hy3?

A context window is not available for the default version.

Is Hy3 open weight?

Yes. The default version is marked as open weight under Apache 2.0.

How many API providers offer Hy3?

15 provider offerings are linked to the default version.

Browse runtime rankings