readable AI benchmarks (simplified)
Tencent
Hy3
Hy3 has a Quality Score of 39.8, ranking 189th among 571 scored models. Per 1M tokens, pricing is $0.136 input, $0.555 output, and $0.034 cached input. It accepts text input, outputs text, and has a 256k token context window. Its Reliability Score is 40.8, ranking 232nd among 573 scored models, above average (total average: 36.5). Its Value Score is 35.6, around average compared with the total Value average of 31.4.
Quality Score39.8coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Score35.6
Factual reliability40.8
Cache Discount75%
Model specification
- Reasoning
- reasoning
- Input modalities
- text
- Output modalities
- text
- Context window
- 256k tokens
- Weights
- open weights
Token prices USD per 1M tokens
- Input
- $0.136
- Output
- $0.555
- Cached Input
- $0.034
- Cache write
- not available
- Cost per task
- $0.072
Capability
- Intelligence
- 25.3
- Coding
- not available
- Agentic
- not available
- Omniscience
- -18.5
- Correct
- 32.0%
- Blended price
- $0.107
Answer outcomes
Fully graded outcomes (Correct + Incorrect): 82.4%
Correct 32.0%Incorrect 50.4%Partial / not attempted 17.6%
Artificial Analysis benchmarks
- GDPval-AA v2
- 28%
- τ³-Banking
- 23%
- SciCode
- 49%
- Humanity’s Last Exam
- 33%
- GPQA Diamond
- 90%
- CritPt
- 5%
- AA-Omniscience
- 41%
- AA-LCR
- 79%
Similar models
- Hy3-preview (Reasoning)TencentQuality Score 33.7Value Score 31.7
- Hy3-preview (Non-reasoning)TencentQuality Score 29.4Value Score 27.6
- Apodex 1.1ApodexQuality Score 39.8Value Score 31.3
- MiMo-V2-Omni-0327XiaomiQuality Score 40.5Value Score not available
- GPT-5.6 Luna (Medium)OpenAIQuality Score 41.0Value Score 35.0
- Agnes 2.5 Pro Alpha (based on Qwen3.5 397B A17B)Sapiens AIQuality Score 39.3Value Score 33.4