---
id: "tencent/hy3"
also: ["tencent/hy3-20260706", "tencent/Hy3", "hf/tencent/Hy3", "requesty/tencent/hy3", "vercel/tencent/hy3"]
card_usd_per_m: "0.132 in, 0.528 out, at tencent/fp8"
cheapest_list_usd_per_m: "0.105 in, 0.435 out, at deepinfra/fp4 on openrouter"
cheapest_effective_usd_per_m: "0.0415 in, 0.528 out, at tencent on requesty"
context: 262144
max_output: 128000
input: ["text"]
output: ["text"]
released: "2026-07-06"
reasoning: ["high", "low", "none"]
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "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"]
endpoints: 17
gateways: "openrouter (8), vercel (5), hf (2), requesty (2)"
no_training: 10
trains: 3
retains_prompts: 7
---
# Tencent: Hy3

Hy3 is built for real world business scenarios with a 295B total, 21B active MoE architecture, native 256K context support, and three reasoning modes. It enhances coding, long form comprehension, multi turn dialogue, and agentic task execution, balancing reliability, efficiency, and cost across high frequency interactions and complex workflows.