Hunyuan Hy3 API — pricing and access
Généraliste et agentic, excellent rapport perf/prix
- Input
- $0.165 USD / 1M tokens
- Output
- $0.682 USD / 1M tokens
- Context
- 256k
- Available on
- Starter and above
- Provider
- Tencent
Try Hunyuan Hy3 in the playground
Hunyuan Hy3 is the generalist agentic model on the Starter plan: it's the one to try first for an assistant that needs to both converse and trigger tool calls — search, API calls, reading files — without moving to a pricier plan. Its 256k window and low price also make it a solid choice for ingesting large documents.
Against DeepSeek V4 Flash, its closest neighbor in price and available at the same Starter tier, Hy3 handles multi-step tasks better, while DeepSeek V4 Flash stays stronger on short, one-off answers. Against MiniMax M3, which targets the same agentic use case but at the Dev tier, Hy3 remains the entry-level pick before moving up if task complexity grows.
It is not the right choice for flagship-grade reasoning — deep legal analysis, math, critical code review — where DeepSeek V4 Pro or Qwen3.7 Max bring a reliability Hy3 doesn't offer at that level of complexity.
Measured on Velqa
- Time to first token
- 0.63 s
- Throughput
- 48.4 tokens/s
Median of 5 streamed calls against api.velqa.dev, 2026-08-10.
Call it
OpenAI-compatible endpoint — change the base URL and the model id, nothing else.
curl https://api.velqa.dev/v1/chat/completions \
-H "Authorization: Bearer $VELQA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "hy3",
"messages": [{"role": "user", "content": "Hello"}]
}'Related models
FAQ
- How much does Hunyuan Hy3 cost on Velqa?
- Hunyuan Hy3 costs $0.165 per million input tokens and $0.682 per million output tokens, in USD. It is included from the Starter plan up.
- Can I use Hunyuan Hy3 without a subscription?
- Yes. Hunyuan Hy3 can be called with prepaid Boost credit, without a monthly plan. Trying it in the playground does need an account, though: anonymous playground access is limited to the demo model.
- Is Hunyuan Hy3 compatible with the OpenAI API?
- Yes. Point your client at https://api.velqa.dev/v1 and set the model to "hy3" — Hunyuan Hy3 answers on the standard chat completions endpoint, streaming included.
