OpenRouter adds Tencent Hunyuan model with 262K context window
TL;DR
OpenRouter adds Tencent Hunyuan Hy3 preview, a mixture-of-experts model with 262K context window. Pricing is $0.07/M input tokens and $0.26/M output tokens with configurable reasoning levels.
What changed
Tencent released Hy3 preview on OpenRouter with 262k context length. Pricing sits at $0.07 per million input tokens and $0.26 per million output tokens. Developers, Basic Users, and Vibe Builders now access its Mixture-of-Experts design for agentic workflows via configurable reasoning modes: disabled, low, or high.
Why it matters
Developers deploying production agents save on costs compared to Llama 3.1 405B, which runs $0.88 per million input tokens on OpenRouter, while Hy3 handles 262k context for long-running tasks. Basic Users run efficient queries without high latency. Vibe Builders adjust reasoning levels to match creative prompts.
What to watch for
Compare Hy3 output quality to DeepSeek-V3 on OpenRouter agent benchmarks. Developers test a 200k token multi-turn conversation on OpenRouter to confirm context retention. Vibe Builders monitor reasoning mode stability in iterative vibe generation.
Who this matters for
- Vibe Builders: Toggle between reasoning modes to match the complexity of your creative outputs.
- Developers: Swap Llama 3.1 405B for Hy3 to slash production costs while maintaining 262k context capacity.
Harsh’s take
Tencent is aggressively undercutting the market with Hy3. By pricing at a fraction of Llama 3.1 405B, they force a choice between brand loyalty and raw operational efficiency. The 262k context window makes this a viable candidate for long-form agentic tasks that previously required expensive proprietary models.
If the reasoning quality holds up, the cost advantage becomes impossible to ignore for any serious production pipeline. However, the model remains a preview. Reliability in agentic workflows is not just about token price, it is about consistent logic across long context windows.
Developers must stress test the reasoning modes against existing benchmarks before migrating critical infrastructure. If Hy3 fails to maintain coherence at the 200k token mark, the low price tag will not save it from being relegated to a secondary utility tool.
by Harsh Desai
About OpenRouter
View the full OpenRouter page →All OpenRouter updatesMore from OpenRouter
- LaunchInclusionAI launches free Ring-2.6-1T (262k context) on OpenRouter
InclusionAI launches free Ring-2.6-1T on OpenRouter. The 1T-parameter-scale model uses 63B active parameters and supports coding agents and tool use.