Tencent's next-generation open-source flagship MoE model, released August 28, 2026. It carries 770B total parameters (49B active per token) with a 1M-token context window under the Apache 2.0 license. Built on a 78-layer mixture-of-experts design (256 routed experts plus one shared expert, top-8 active per token), it roughly doubles both capacity and context versus Hy3 (295B/21B, 256K). It scores 85.4 on Terminal-Bench 2.1, 74.1 on Toolathlon-Verified, and 37.1 on APEX-Agents (pass@1), positioning it as a productivity-focused model for software engineering, office analytics, game development, and scientific research. Weights ship on Hugging Face and ModelScope, with API access via Tencent Cloud TokenHub and OpenRouter at $0.834 input / $2.501 output per 1M tokens (cache hits $0.042).
Parameters
770B
Context Window
1M
License
Apache 2.0
Release Date
2026-08-28
API Pricing
Input Price (per 1M tokens)
$0.834
Output Price (per 1M tokens)
$2.501
Billing Mode: standard
Strengths
- •Open-weight flagship at 770B total / 49B active parameters
- •1M-token context handles long-horizon and long-document tasks
- •85.4 on Terminal-Bench 2.1 signals Opus 5-tier agentic coding
- •Apache 2.0 lets teams self-host, fine-tune, and ship commercially
Weaknesses
- •No vision or other multimodal capability — text only
- •Tends to over-think and over-verify, raising latency and token use
- •Explicitly an early preview with headroom in pre- and post-training
- •Very large checkpoint raises self-hosting memory requirements
Use Cases
- •Long-horizon, multi-file software-engineering agents
- •Complex office document analysis (finance, insurance)
- •Conversational game prototyping with Unity or Unreal
- •Scientific research support (molecular dynamics, foundational math)
Deep Analysis
Total / Active Parameters
770B / 49B (MoE)
78 layers; 256 routed + 1 shared expert, top-8 active; 1M-token context
Context Window
1,048,576 tokens
4x the 256K of Hy3; native MTP speculative-decoding layer
License
Apache 2.0
Weights on Hugging Face and ModelScope; commercial use permitted
Input / Output Price
$0.834 / $2.501 per 1M
Cache hits $0.042/1M; a fraction of closed frontier cost
Terminal-Bench 2.1
85.4
14.6 pts above Hy3; Tencent claims parity with Claude Opus 5
SWE-Bench Pro
51.2
DeepSWE 64.3; Toolathlon-Verified 74.1; APEX-Agents 37.1
Strengths
- ・Open-weight flagship at 770B/49B that roughly doubles Hy3's capacity and context, landing in the top tier of permissively-licensed models.
- ・85.4 on Terminal-Bench 2.1 and 74.1 on Toolathlon-Verified put it at Opus 5-tier on agentic coding, per Tencent's internal blind test (2.99/4.00 vs GLM-5.3 2.92, Kimi K3 2.94).
- ・Apache 2.0 weights with official vLLM/SGLang recipes and API access via TokenHub and OpenRouter at $0.834/$2.501 per 1M tokens.
- ・Self-improving R&D loop: Tencent reports Hy4 helped optimize its own inference stack for a 31.8% end-to-end throughput gain.
Weaknesses
- ・Text-only — no vision or other multimodal capability, unlike GLM-5.3 or DeepSeek V4-Flash-Vision-Exp.
- ・Tends to over-think and over-verify, which can raise latency and token consumption on complex tasks.
- ・Explicitly an early preview; Tencent states pre-training and post-training both have substantial headroom.
- ・The full checkpoint is very large, so self-hosting still demands serious memory and serving engineering.
Competitor Comparison
| Model | Arena | SWE | GPQA | Price |
|---|---|---|---|---|
| Tencent Hy4 preview | 85.4 (Terminal-Bench 2.1) | 51.2 (SWE-Bench Pro) | N/A | $0.834/$2.501 per 1M |
| GLM-5.3 | ~84.5 (Terminal-Bench 2.1) | N/A | N/A | MIT; OpenRouter |
| Kimi K3 | ~83 (Terminal-Bench 2.1) | N/A | N/A | Open weight |
| DeepSeek V4 Pro | ~85 (Terminal-Bench 2.1) | N/A | N/A | Closed API |
Tencent Hy4 preview is the company's largest permissively-licensed model to date — a 770B/49B MoE with a 1M-token context, open-sourced August 28, 2026 under Apache 2.0. It is built for real productivity work (coding, office analytics, game dev, science) and posts Opus 5-adjacent agentic benchmarks at a fraction of frontier cost.
Sources
Analysis generated: 2026-09-02