Every Tencent model behind one OpenAI-compatible endpoint. You pay Tencent's published token rate — OrcaRouter adds $0. Below: live input, output and cached-input pricing per million tokens, context and output limits, measured latency, and the code to call any of them without changing your SDK.
| Model | Input per million tokens | Output per million tokens | Cached in per million tokens | Context | Max out | Quality | Try |
|---|---|---|---|---|---|---|---|
| Hy3 | $0.180 | $0.590 | $0.059 | 262K | — | 8 | Try |
| Hy4 preview | $0.834 | $2.50 | $0.042 | 1M | 64K | 8 | Try |
| Hy3 (Free) | $0.00 per request | 262K | — | — | Try | ||
| Hy4 preview (Free) | $0.00 per request | 1M | 64K | — | Try | ||
Estimate based on list price
Estimating cost for Hy3: $3.03 per month
Estimates use the provider's published rates. OrcaRouter adds $0 markup, so this is what you would pay calling them directly.
curl https://api.orcarouter.ai/v1/chat/completions \
-H "Authorization: Bearer $ORCAROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "tencent/hy3",
"messages": [{"role": "user", "content": "Hello"}]
}'Already using the Tencent SDK? Point base_url at OrcaRouter and keep everything else.
Exactly the same as calling Tencent directly. Hy3 is $0.18 per million input tokens and $0.59 per million output tokens; the cheapest Tencent model here is Hy3 at $0.18 per million input. OrcaRouter adds $0 markup — you are billed the provider's published rate.
Yes. Add your own Tencent key in the console and OrcaRouter routes through it, so usage bills to your account with the provider while you keep the unified endpoint, logging, spend controls and automatic failover.
No. Point base_url at https://api.orcarouter.ai/v1, keep the OpenAI-compatible client you already use, and reference the model as tencent/<model>. Nothing else in your code changes.
Hy4 preview, at 1048K tokens. The full context and max-output limits for every Tencent model are in the pricing table above.
Every model page carries side-by-side comparisons against the closest models from other providers on price, context, latency and benchmarks. Because OrcaRouter adds no markup, the prices you are comparing are the providers' own.
Across the Tencent catalogue: Tools (4), Reasoning (4), JSON mode (4). Use the capability chips above the pricing table to narrow the list to the models that support what you need.