Tencent Hy
August 28, 2026
Introducing Hy4 preview
Today we're releasing Hy4 preview, our most capable model to date. It's a 770B-parameter model with 49B active parameters and a 1M-token context window, and it makes the biggest gains where the work is hardest: long-horizon software engineering, document-heavy office work, and scientific research.
[](https://hy.tencent.ai/research/hy4-preview#a-new-flagship-generation)A new flagship generation
We scaled Hy4 preview on three fronts: model size, context length, and training data. Stronger pre-training and a substantially larger post-training run compound into another step change in capability — the largest generation-over-generation gain we've measured, and enough to put Hy4 preview at the open-source frontier.
!Image 2: Hy4 preview-bench.png
[](https://hy.tencent.ai/research/hy4-preview#built-for-productivity)Built for productivity
We partnered with top experts inside Tencent — such as software engineers, game developers, finance analysts, and security experts — and built training data around the work they ship. The result is a model that gets meaningfully further on the tasks these teams run every day:
[](https://hy.tencent.ai/research/hy4-preview#software-engineering)Software engineering
Better at understanding, planning, debugging, and verifying long-horizon development tasks, with further gains in the visual taste and interaction quality of front-end work.
A desktop 3D viewer for seven Chinese historic buildings, styled as a quiet museum research file: fixed three-column layout, rice-paper background, serif type, one indigo accent, hairline rules. Central Three.js stage with drag-rotate and zoom; no dark tech aesthetic, no page scrolling, nothing upstaging the model.
[](https://hy.tencent.ai/research/hy4-preview#office-and-analysis)Office and analysis
Takes messy context spread across many files and converts it into shareable artifacts — documents, spreadsheets, and presentations — handling data analysis, equations, and financial models with greater precision.
Review 24 reimbursement claims: approve, reduce, or return each one. Dig through the policies (several versions, work out which applies), roster, budgets, invoice ledger and emails yourself. Output a summary table, per-claim reasoning citing sources and arithmetic, a list of suspected fraud, and what documents to chase.
[](https://hy.tencent.ai/research/hy4-preview#game-development)Game development
Turns a single prompt into a playable prototype and works fluently with game engines, so developers can keep refining complex projects over multiple turns.
A third-person cartoon 3D game in Unity/Tuanjie: a penguin explores a snowfield, gathers berries, opens chests, fights monsters and escapes. Includes menu/loading/pause/results screens, animation blend tree, inventory, quest HUD, music and sound. All feel-related values exposed as editor parameters.
[](https://hy.tencent.ai/research/hy4-preview#scientific-research)Scientific research
Stronger understanding, reasoning, and problem-solving on hard research questions, with solid progress across AI research, molecular dynamics, condensed matter physics, and pure mathematics.
Hy4 preview managed several Codex sessions in parallel and adjusted its research direction as results came in. On a small-model post-training task it acted as the researcher, coordinating Codex to optimize several evaluation targets at once, and beat Codex exploring on its own across all 8 benchmarks — evidence it can judge direction, organize experiments and iterate on complex R&D work.
We also continue to co-design Hy4 preview with Tencent products like CodeBuddy and WorkBuddy, so that gains in the model show up in the work people actually do with it. To check that, we ran a blind side-by-side evaluation: 163 internal experts rated model outputs on 203 engineering tasks. Hy4 preview came out slightly ahead of both GLM 5.3 (2.99 vs. 2.92 average, 46.8% wins / 12.8% ties / 40.4% losses) and Kimi K3 (2.99 vs. 2.94, 51.2% wins / 7.9% ties / 40.9% losses).
Hy4 preview is open-sourced and available today on Tencent Cloudand OpenRouter, and you can use it now in Tencent products such as CodeBuddy / WorkBuddy, Yuanbao, and ima. We believe in keeping the model open and affordable: that's how progress gets into more hands.
Hy4 preview API Pricing（per 1M tokens）
| Cached Input | Input | Output | | --- | --- | --- | | $0.042 | $0.834 | $2.501 |
This is an early version of Hy4. There is real headroom left in both pre-training and post-training, and we are shipping with known issues — among them, spending longer than necessary reasoning through complex tasks, and a tendency to over-verify its own work. We'll keep iterating quickly on these. As with Hy3 preview, we would rather ship early and hear what breaks — that's what made Hy3 substantially better, and it's how we will get Hy4 right. We will also keep collaborating closely with Tencent's products and in-house experts to push the boundaries of model intelligence while making it more abundant and affordable.
