Tencent has released Hy4, a new model that adds vision capabilities, and is positioning it as a leading open model for finance, frontend development, games, office work, and research.
Tencent · Hy4 release
Tencent's Hy4 pitches a vision-enabled open model for professional work, trailing only the top proprietary systems
Hy4 adds image understanding and targets finance, frontend development, games, office work, and research. Tencent places it among the strongest open models in its peer group, while acknowledging a gap to two proprietary frontier systems.
5
Professional domains Tencent highlights: finance, frontend, games, office, research
3
Competing open models in the comparison: GLM 5.3, DeepSeek V4.1 Flash, Kimi K3.1 Pro
2
Proprietary frontier systems Hy4 still trails on most measures: Claude Opus 5.5, GPT 6.1 Sol
What the next evidence has to show
Three open questions decide whether the release translates into adoption
01
Independent evaluations
→
02
Vision on real-world tasks
→
03
Adoption by developers and businesses
The case for Hy4
Among the strongest open models in its peer group, per Tencent
Adds image input, widening what it can handle versus text-only systems
Broad professional scope gives developers and enterprises more options outside closed APIs
The caveats
Trails both proprietary frontier systems on most evaluated measures
Benchmark results depend on chosen test suites, prompts, and versions, and are not independently verified here
Parameter counts, licensing terms, and per-benchmark scores were not disclosed in the available material
Bottom line
Strong among open models, not yet at the frontier
Hy4's value rests on independent results and real-world vision performance. Until those arrive, its position is a company claim against a clear gap at the top of the market.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...