Skip to contentx402 is live on Robinhood Chain: agents can now pay for any model, per call, in USDG →
OmniRail
← All models

Tencent · Language model

Hy3

tencent/hy3
  • Language
Released Jul 6, 2026

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:...

Specifications

Context window
262,144 tokens
Max output
128,000 tokens
Input
text
Output
text
Reasoning effort
high, low, none
Tokenizer
Other

Supported parameters

Accepted by the model and forwarded as-is by the gateway.

  • frequency_penalty
  • logit_bias
  • max_completion_tokens
  • max_tokens
  • min_p
  • presence_penalty
  • reasoning
  • reasoning_effort
  • repetition_penalty
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_k
  • top_p

Use it

POST /api/v1/chat/completions · MCP chat · full reference

curl https://omnirail.org/api/v1/chat/completions \
  -H "Authorization: Bearer $OMNIRAIL_KEY" -H "Content-Type: application/json" \
  -d '{"model":"tencent/hy3","messages":[{"role":"user","content":"Hello"}],"max_tokens":300}'