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 Oct 23, 2026. Default version: Hy3
LLMBoard Score70.7Hy3
Coverage80%29 benchmark families
Context window262.1KTokens
Official input priceN/AOfficial price unavailable
Hy3 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Hy3 Benchmark Results
Benchmark scores for Hy3.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkCL-bench | Score23.80% | Rank01 | Participants3 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCL-bench (Life) | Score17.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCMT-Benchmark | Score37.90% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkPHYBench | Score77.40% | Rank01 | Participants3 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkHorizonMath | Score7.10% | Rank02 | Participants4 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkFrontierScience Olympiad | Score74.80% | Rank03 | Participants4 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkIMO-AnswerBench | Score90.00% | Rank03 | Participants22 | Percentile90.48% | EvidenceC | Evaluated |
| BenchmarkUSAMO 2026 | Score30.24 points | Rank03 | Participants3 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkClaw-Eval | Score68.50% | Rank04 | Participants14 | Percentile76.92% | EvidenceC | Evaluated |
| BenchmarkFrontierScience Research | Score21.30% | Rank04 | Participants5 | Percentile25.00% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam (no tools, text-only) | Score47.00% | Rank04 | Participants14 | Percentile76.92% | EvidenceC | Evaluated |
| BenchmarkSuperChem | Score54.90% | Rank04 | Participants5 | Percentile25.00% | EvidenceC | Evaluated |
| BenchmarkWildClawBench | Score53.60% | Rank04 | Participants6 | Percentile40.00% | EvidenceC | Evaluated |
| BenchmarkAA-LCR | Score73.40% | Rank05 | Participants27 | Percentile84.62% | EvidenceC | Evaluated |
| BenchmarkArXivMath | Score52.20% | Rank05 | Participants5 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkDeepSearchQA | Score91.00% | Rank05 | Participants13 | Percentile66.67% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam (with tools, text-only) | Score53.20% | Rank06 | Participants10 | Percentile44.44% | EvidenceC | Evaluated |
| BenchmarkSkillsBench | Score55.30% | Rank06 | Participants12 | Percentile54.55% | EvidenceC | Evaluated |
| BenchmarkWideSearch | Score76.40% | Rank06 | Participants15 | Percentile64.29% | EvidenceC | Evaluated |
| BenchmarkMathArena Apex | Score38.70% | Rank07 | Participants11 | Percentile40.00% | EvidenceC | Evaluated |
| BenchmarkAPEX-Agents | Score25.60% | Rank11 | Participants11 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkMCP Atlas | Score79.10% | Rank11 | Participants39 | Percentile73.68% | EvidenceC | Evaluated |
| BenchmarkDeepSWE | Score28.00% | Rank12 | Participants13 | Percentile8.33% | EvidenceC | Evaluated |
| BenchmarkBrowseComp | Score84.20% | Rank16 | Participants70 | Percentile78.26% | EvidenceC | Evaluated |
| BenchmarkNL2Repo | Score45.60% | Rank16 | Participants25 | Percentile37.50% | EvidenceC | Evaluated |
| BenchmarkSWE-bench Multilingual | Score75.80% | Rank17 | Participants50 | Percentile67.35% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score78.00% | Rank24 | Participants120 | Percentile80.67% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Steerability | Score-0.37% | Rank24 | Participants44 | Percentile46.51% | EvidenceA | Evaluated |
| BenchmarkToolathlon | Score48.50% | Rank26 | Participants42 | Percentile39.02% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score90.40% | Rank27 | Participants256 | Percentile89.80% | EvidenceC | Evaluated |
Hy3 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenaagent steerability | Categoryoverall | Rank24 | Rating / score0.00 | VotesN/A | Observations40.5K | Result date |
| Arenatext style control | Categoryindustry mathematical | Rank24 | Rating / score1,492.53 | Votes754 | ObservationsN/A | Result date |
| Arenatext style control | Categoryspanish | Rank26 | Rating / score1,472.44 | Votes458 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry mathematical | Rank29 | Rating / score1,482.21 | Votes754 | ObservationsN/A | Result date |
| Arenatext style control | Categorymath | Rank30 | Rating / score1,481.02 | Votes689 | ObservationsN/A | Result date |
| Arenaagent praise complaint | Categoryoverall | Rank32 | Rating / score-0.07 | VotesN/A | Observations5K | Result date |
| Arenatext | Categorymath | Rank32 | Rating / score1,473.02 | Votes689 | ObservationsN/A | Result date |
| Arenaagent | Categoryoverall | Rank33 | Rating / score-0.06 | VotesN/A | Observations1.2M | Result date |
| Arenatext factuality | Categoryindustry mathematical | Rank33 | Rating / score1,465.03 | Votes671 | ObservationsN/A | Result date |
| Arenatext style control | Categorygerman | Rank35 | Rating / score1,467.34 | Votes227 | ObservationsN/A | Result date |
| Arenaagent task outcome explicit | Categoryoverall | Rank36 | Rating / score-0.11 | VotesN/A | Observations13.9K | Result date |
| Arenatext style control | Categorypolish | Rank36 | Rating / score1,463.13 | Votes241 | ObservationsN/A | Result date |
| Arenaagent bash recovery steps | Categoryoverall | Rank38 | Rating / score-0.07 | VotesN/A | Observations29.1K | Result date |
| Arenatext | Categoryspanish | Rank38 | Rating / score1,456.83 | Votes458 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry life and physical and social science | Rank40 | Rating / score1,482.77 | Votes2,014 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-react | Rank40 | Rating / score1,510.83 | Votes3,947 | ObservationsN/A | Result date |
| Arenatext | Categorypolish | Rank41 | Rating / score1,446.56 | Votes241 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry life and physical and social science | Rank41 | Rating / score1,478.79 | Votes1,991 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry medicine and healthcare | Rank41 | Rating / score1,480.53 | Votes821 | ObservationsN/A | Result date |
| Arenatext style control | Categorychinese | Rank42 | Rating / score1,501.33 | Votes973 | ObservationsN/A | Result date |
| Arenaagent bash reliability | Categoryoverall | Rank42 | Rating / score-0.12 | VotesN/A | Observations10.4K | Result date |
| Arenaagent tool hallucination | Categoryoverall | Rank43 | Rating / score-0.04 | VotesN/A | Observations1.1M | Result date |
| Arenaagent bash reliability no prereq | Categoryoverall | Rank43 | Rating / score-0.14 | VotesN/A | Observations7.3K | Result date |
| Arenatext | Categorygerman | Rank44 | Rating / score1,448.51 | Votes227 | ObservationsN/A | Result date |
| Arenawebdev | Categoryoverall | Rank44 | Rating / score1,507.10 | Votes6,256 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev | Rank44 | Rating / score1,507.10 | Votes6,256 | ObservationsN/A | Result date |
| Arenatext style control | Categoryexpert | Rank45 | Rating / score1,489.87 | Votes1,358 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-html | Rank46 | Rating / score1,486.65 | Votes970 | ObservationsN/A | Result date |
| Arenatext style control | Categoryfrench | Rank48 | Rating / score1,479.17 | Votes521 | ObservationsN/A | Result date |
| Arenatext style control | Categoryinstruction following | Rank48 | Rating / score1,447.29 | Votes4,617 | ObservationsN/A | Result date |
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
- 22
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderTencent TokenHub | Provider model IDhy3 | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context256K | Updated |
| ProviderTencent Token Plan | Provider model IDhy3 | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context256K | Updated |
| ProviderNanoGPT | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.066 | Output / 1M$0.26 | Context262.1K | Updated |
| ProviderOpenRouter | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.0825 | Output / 1M$0.33 | Context262.1K | Updated |
| ProviderKilo Gateway | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.0825 | Output / 1M$0.33 | Context262.1K | Updated |
| ProviderDeep Infra | Provider model IDtencent/Hy3 | Regionglobal | Input / 1M$0.13 | Output / 1M$0.53 | Context262.1K | Updated |
| ProviderDevPass (LLM Gateway) | Provider model IDhy3 | Regionglobal | Input / 1M$0.132 | Output / 1M$0.528 | Context262.1K | Updated |
| ProviderLLM Gateway | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.132 | Output / 1M$0.528 | Context262.1K | Updated |
| ProviderSiliconFlow | Provider model IDtencent/Hy3 | Regionglobal | Input / 1M$0.132 | Output / 1M$0.528 | Context262.1K | Updated |
| ProviderEden AI | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.132 | Output / 1M$0.528 | Context256K | Updated |
| ProviderJalapeno Cloud | Provider model IDHy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context202.8K | Updated |
| ProviderHugging Face | Provider model IDtencent/Hy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context262.1K | Updated |
| ProviderOpenCode Go | Provider model IDhy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context256K | Updated |
| ProviderRequesty | Provider model IDhy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context262.1K | Updated |
| ProviderVercel AI Gateway | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context262.1K | Updated |
| ProviderNovita AI | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context262.1K | Updated |
| ProviderCommandCode | Provider model IDtencent/hy3-paid | Regionglobal | Input / 1M$0.14 | Output / 1M$0.58 | Context256K | Updated |
| ProviderKenari | Provider model IDhy3 | Regionglobal | Input / 1M$0.1543 | Output / 1M$0.6286 | Context256K | Updated |
| ProviderAIHubMix | Provider model IDhy3 | Regionglobal | Input / 1M$0.1562 | Output / 1M$0.6248 | Context256K | Updated |
| ProviderCrossModel | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.16 | Output / 1M$0.64 | Context262.1K | Updated |
| ProviderOrcaRouter | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.18 | Output / 1M$0.59 | Context256K | Updated |
| ProviderAI/ML API | Provider model IDtencent/hy3 | Regionglobal | Input / 1M$0.2116 | Output / 1M$0.8464 | Context262.1K | Updated |
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
- 262.1K
- Max output
- 262.1K
- 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.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionHy3 | Released | LLMBoard70.68 | Parameters295B | Context262.1K | Max output262.1K | Open weightsYes | LicenseApache 2.0 |
Models similar to Hy3
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
TE
Hy4
Tencent
AN
Claude Sonnet 4.6
Anthropic
OP
GPT-6-Luna
OpenAI
XA
Grok 4.7
xAI
BY
Seed 2.1 Turbo
ByteDance
AC
Qwen3.7 Plus
Alibaba Cloud / Qwen Team
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 2026-10-22.
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?
The default version has a 262.1K token context window.
Is Hy3 open weight?
Yes. The default version is marked as open weight under Apache 2.0.
How many API providers offer Hy3?
22 provider offerings are linked to the default version.