AI model
tencent/hy3-reasoning
Tencent: Hy3 (Reasoning) is a 262,144-token model served on Composite at $0.103 per million tokens in and $0.413 per million tokens out.
Try in Playground| Model ID | tencent/hy3-reasoning |
|---|---|
| Provider | cv11 |
| Context window | 262,144 tokens |
| Accepts | text |
| Returns | text |
| Input price | $0.103 per million tokens |
| Output price | $0.413 per million tokens |
| Prompt caching | Supported — cached input is billed at the provider's cache rate, on credits and plans alike |
| Thinking | Optional - see "Tencent: Hy3 (Reasoning) (Reasoning)" in the catalog to turn it on by default |
| Moderated | No |
| Cost on a monthly plan | About 1.0 plan credits per typical call (8k in, 1k out) |
| Routing endpoints | 1 |
Make a Composite API key and type tencent/hy3-reasoning as the model name. JanitorAI takes the full URL https://composite.lucidity.sh/v1/chat/completions; SillyTavern, ChubAI, Agnai and RisuAI take https://composite.lucidity.sh/v1. It runs on credits, a plan, or your own provider key.
JanitorAI setup · SillyTavern setup · Best Janitor AI proxies compared
This model has not carried enough traffic on Composite yet to publish a success rate, a latency figure or a traffic rank. The numbers appear here once it has.
$0.103 per million tokens input · $0.413 per million tokens output
Credits pay the provider’s price plus 25%; cache hits cost less.
On a monthly plan, a typical call (8k tokens in, 1k out) uses about 1.0 plan credits, at the provider's own price.
cv11
This model is available on Composite alongside dozens of others from other labs, so it is easy to compare against similar options.
With a 262,144-token context window, it can hold a long-running roleplay or a full lorebook in memory without losing earlier plot details, so it suits multi-session campaigns and character cards with heavy backstory.
It is priced low enough for high-volume daily roleplay without the per-message cost adding up quickly.
Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...
Supplied by cv11, not written by Composite.
Composite Anthropic Router · Composite Gemini Router · Composite GPT Router