Hy3是腾讯混元的生產級混合專家模型(Mixture-of-Experts)——總參數295B,每次傳遞僅有21B活躍參數(192個專家,top-8路由),是基於Hy3-preview系列升級的版本。它擴大了RL訓練規模和訓練後資料質量,以在推理、長上下文和智能體任務上進一步提升,達到與參數規模數倍於它的旗艦模型相當的結果。它提供256K token的上下文窗口(文本輸入,文本輸出),並具有可配置的推理努力程度,專為真實世界的編碼、工具使用和多步驟智能體工作流程而設計,在強品質成本比下運作。
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.orcarouter.ai/v1",
api_key=os.environ["ORCAROUTER_API_KEY"],
)
response = client.chat.completions.create(
model="tencent/hy3-free",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)frequency_penaltyinclude_reasoninglogit_biasmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_p本週開發者的討論
@misc{orcarouter_hy3_free,
title = {Hy3 (Free) API},
author = {Tencent},
year = {n.d.},
howpublished = {OrcaRouter},
url = {https://www.orcarouter.ai/models/tencent/hy3-free}
}Tencent. (n.d.). Hy3 (Free) API. OrcaRouter. https://www.orcarouter.ai/models/tencent/hy3-free