Tencent released and open-sourced Hy4 preview, its next-generation LLM: a 770B-parameter MoE with 49B active per token and a context window beyond 1M tokens, shipped as standard and FP8 weights under Apache 2.0 (per HF tags; the announcement itself names no license). Target uses are software engineering (planning, debugging, front-end quality), office work (financial and data analysis, cross-document collaboration), game development (playable prototypes from a single prompt) and scientific research. Public benchmark evidence is thin: one internal Tencent blind test with 163 experts over 203 engineering tasks scored Hy4 preview 2.99/4.00, slightly ahead of GLM-5.3 (2.92) and Kimi K3 (2.94). Tencent says the model assisted its own training pipeline and self-optimized its inference stack for a 31.8% end-to-end throughput gain. The full Hy4 is not released yet — a preview-first rollout with the next batch 'soon'. API access via Tencent Cloud TokenHub and OpenRouter at $0.834/M input, $2.501/M output, $0.042/M cached.
Tencent open-weights Hy4 preview: 770B MoE under Apache 2.0
Tencent released Hy4 preview as a 770B-parameter MoE with 49B active parameters per token and a context window beyond 1M tokens. Standard and FP8 weights use Apache 2.0. Tencent's announcement reports an internal expert blind test and a 31.8% inference-throughput gain from model-assisted optimization. The Hugging Face page now exposes public leaderboard metadata, including SWE-bench Pro and Terminal-Bench 2.1 scores, but it does not identify a reproducible run or an independent evaluator. The full Hy4 is still absent; only the preview and FP8 repositories are listed.
Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.
What this means for you
For teams evaluating open weights, Hy4 now has stronger adoption evidence: the main repository rose from 8,219 to 22,788 monthly downloads since the September 11 check, despite its 780B scale. That is meaningful distribution, but not independent capability validation. Keep it at WATCH until a third party publishes a reproducible run against the public weights or Tencent ships the full model.
Evidence & related material
As of 2026-09-27, the Hy4 preview main repository records 22,788 monthly downloads, up 177% from the September 11 check; the full model remains unreleased.
Subsequent changes
Hy4 preview reached 22,788 monthly downloads, but the full model and reproducible independent evaluation remain absent.
Technical details & scores
| Architecture | 770B total / 49B active MoE; context beyond 1M tokens |
|---|---|
| License | Apache 2.0 per Hugging Face tags (announcement text names no license) |
| Artifacts | tencent/Hy4-preview + tencent/Hy4-preview-FP8 on HF (repos created 2026-08-27T08:52Z); ModelScope mirror; vLLM recipe day-0; downloads ~1.4k main + ~1.3k FP8, 283 likes @2026-08-30 |
| Eval | Tencent internal blind test: 163 experts, 203 engineering tasks; Hy4 preview 2.99/4.00 vs GLM-5.3 2.92 and Kimi K3 2.94. HF displays leaderboard metadata including SWE-bench Pro 65.7 and Terminal-Bench 2.1 85.4, but provides no reproducible run or named independent evaluator. |
| Self Improvement | model assisted its own training pipeline; self-optimized inference system +31.8% end-to-end throughput (vendor claim) |
| API | Tencent Cloud TokenHub + OpenRouter; $0.834/M input, $2.501/M output, $0.042/M cached; free two weeks on WorkBuddy/CodeBuddy; Hy3 free extended to 2026-09-30 |
| Rollout | preview-first; full Hy4 weights not yet released; next Hy4-series batch 'expected soon' |
Historical update notes
· Adoption & follow-up check @2026-08-31T00:00Z: downloads 2,123 main (+~700 vs the 8/30 sample) + 319 likes; FP8 1,469 — modest velocity compared with the GLM-5.3-Flash cohort. No full Hy4 release yet; the announcement still only says the next batch is expected 'soon' (community speculates ~2.5 months to GA from the Hy3 precedent). Verification note: aggregators (datalearner etc.) list public benchmark scores (Terminal Bench 2.1 85.4, SWE-Bench Pro 65.7) that do NOT appear in the official announcement or the HF model card — unverified, not counted; the official page still shows only the internal blind test. Community GGUF compression to ~200GB claimed at ~98% performance retention (unverified). Recommendation stays WATCH.
· adoption check @2026-09-02T00:0xZ: 3,516 downloads (+66% vs 2,123 @8/31) + 383 likes — continued but still modest vs the GLM-5.3-Flash cohort; full Hy4 not shipped ('soon' unchanged), no official public benchmarks. Trend #1 criterion (b) candidate status unchanged. Stays WATCH
· Adoption check @2026-09-05: 5,684 downloads (+62% vs 9/2) + 430 likes; FP8 2,666. Full Hy4 still not shipped (tencent org's latest repos remain Hy4-preview/Hy4-preview-FP8 from 8/27; flagship line still Hy3 from July). Third-party quantization ecosystem growing despite the preview: AngelSlim/Hy4-preview-GGUF 109,240 downloads; mlx-community 4bit; inferencerlabs MLX Q4i (9/2); anemll FlashMoE-STQ1_0 (9/3). Stays WATCH