Hy3
Tencent · released
Tencent's smaller generally available open-weight model, priced as a low-cost workhorse for agents and production applications.
Best for: Low-cost text workloads · Apache-licensed open weights
- Intelligence Index
- 25.3
- #29 of 38
- Price per 1M tokens
- $0.13 in $0.53 out
- Blended $0.23 at 3:1
- Context window
- 256K
- Up to 128K output
- Output speed
- 85 tokens/s
- $0.07 measured cost per task
Capability profile
One 0–100 score per area, placing this model between the weakest and strongest tracked model across that area's benchmarks.
- Overall41 · #26 of 38
Based on 2 of 3 benchmarks
- Coding23 · #28 of 39
Based on 3 of 5 benchmarks
- Agents & knowledge work36 · #28 of 40
Based on 2 of 6 benchmarks
- Reasoning & science31 · #29 of 39
Based on 2 of 2 benchmarks
- Knowledge & reliability49 · #31 of 39
Based on 2 of 2 benchmarks
- VisionNot enough data
Based on 0 of 2 benchmarks
Benchmark results
All third-party results on record. Rank is among the tracked models that have been tested on that benchmark.
Overall
#29 of 38· Artificial Analysis
- Vals IndexNot tested
- Arena · Text1456 ±7
#18 of 33· 10,476 votes· Arena
Coding
#31 of 38· Artificial Analysis
- SciCode48.6%
#26 of 38· Artificial Analysis
- Arena · WebDev1508 ±9
#25 of 34· 5,543 votes· Arena
- Vibe Code BenchNot tested
- Code MigrationNot tested
Agents & knowledge work
- GDPval-AA28.3
#28 of 39· Artificial Analysis
- Arena · Agent-5.4 ±1.2
#24 of 29· Arena
- Finance Agent v2Not tested
- Excel Modeling BenchmarkNot tested
- Legal Research BenchNot tested
- Tax Agent BenchNot tested
Reasoning & science
- Humanity's Last Exam33.5%
#28 of 39· Artificial Analysis
- CritPt4.9%
#29 of 39· Artificial Analysis
- ARC-AGI-2saturatingNot tested
- ARC-AGI-3Not tested
- GPQA Diamondsaturating89.7%
#15 of 27· Artificial Analysis
Knowledge & reliability
#33 of 39· Artificial Analysis
- AA-LCR79.0%
#28 of 39· Artificial Analysis
Vision
- MMMU-ProNot tested
- Arena · VisionNot tested
Specifications
- Lab
- Tencent
- Released
- Jul 6, 2026
- Availability
- Generally available
- Context window
- 256,000 tokens
- Max output
- 128,000 tokens
- Accepts
- Text
- Knowledge cutoff
- Not published
- Weights
- Open
- Licence
- Apache 2.0
- Parameters
- 295B total · 21B active
- Reasoning control
- Effort off, low or high
- API model ID
hy3
Pricing
- Input / 1M tokens
- $0.13
- Cached input / 1M
- $0.03
- Output / 1M tokens
- $0.53
- Blended (3:1)
- $0.23
- Measured cost / task
- $0.07
Worth knowing
- Release date is the first public listing; Tencent published no dated announcement.
Official sources
Closest alternatives
The models nearest to this one on overall score.
23 benchmarks tracked. See methodology and sources.