Hy3

tencent/hy3
NewFeatured
ToolsJSONReasoning
by Tencent · 2026-07-06

Hy3 is Tencent Hunyuan's production-grade Mixture-of-Experts model — 295B total parameters with only 21B active per pass (192 experts, top-8 routing), the upgraded release built on the Hy3-preview line. It expands RL-training scale and post-training data quality for further gains in reasoning, long-context, and agentic tasks, reaching results comparable to flagship models several times its parameter size. It serves a 256K-token context window (text in, text out) with configurable reasoning effort, and is built for real-world coding, tool use, and multi-step agent workflows at a strong quality-to-cost ratio.

ctx262.1K tokens
Inputtext
Outputtext
p50 TTFT3.77 s
INPUT$0.18/ 1M tokens
OUTPUT$0.59/ 1M tokens
p50 TTFT3.77 s7d
p95 TTFT7.45 s7d
TRAFFIC305.7Ktokens / 7d

Tencent Hy3 is a text-only large language model developed by Tencent. It is designed to process and generate text with a context window of 262,144 tokens, enabling it to work with very long documents…

What is Tencent Hy3?

Who would benefit from using Tencent Hy3?

What are the key technical specifications?

How does the pricing model work?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="tencent/hy3",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • frequency_penalty
  • include_reasoning
  • logit_bias
  • 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

Pricing

Input / 1M tokens$0.180
Output / 1M tokens$0.590
Cache read / 1M$0.059
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $3.03 · With prompt caching $2.61

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.000299

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
3.77 s
Output speed
89.0 tok/s
p95 TTFT
7.45 s
Error rate
0%

Public benchmarks

AA-LCR
73.4
BrowseComp
84.2
ClawEval (pass^3)
68.5
FrontierScience-Olympiad
74.8
HLE (with tools, text-only)
53.2
MathArena Apex
38.7
MCP Atlas (public)
79.1
NL2repo
45.6
SkillsBench (text-only)
55.3
SWE-bench Multilingual
75.8
SWE-bench Pro
57.9
Terminal Bench 2.1
71.7
Source: tencent

FAQ

What is the cost of using Tencent Hy3?
$0.18 per 1 million input tokens and $0.59 per 1 million output tokens, billed at the provider's exact rate with zero markup via OrcaRouter.
What is the context window size?
262,144 tokens.
What are the main strengths of Tencent Hy3?
Its very large context window (262k) and strong performance on the BrowseComp benchmark (84.2) make it ideal for long-document analysis. The pricing is competitive for its context size.
How does Tencent Hy3 compare to smaller, cheaper models?
It is better for tasks requiring a large context without chunking, but for short inputs, smaller models are faster and cheaper.
Does OrcaRouter share my data with Tencent?
Data handling policies are specified in OrcaRouter's terms of service; no specific sharing details are provided for this model.
How do I call Tencent Hy3 using the OpenAI Python client?
Set openai.api_base = 'https://api.orcarouter.ai/v1', openai.api_key = 'your-key', and use model='tencent/hy3'.
Can I stream responses from Tencent Hy3?
Yes, set the 'stream' parameter to true in the API call following standard OpenAI streaming.
What is the maximum output length?
You can set max_tokens up to the remaining context window after input, but no specific output limit is advertised.

Embed this badge

Tencent: Hy3$0.18/M in3772ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/tencent/hy3" target="_blank"> <img src="https://www.orcarouter.ai/embed/tencent/hy3.svg" alt="Tencent: Hy3 on OrcaRouter" /> </a>
Markdown [![Tencent: Hy3](https://www.orcarouter.ai/embed/tencent/hy3.svg)](https://www.orcarouter.ai/models/tencent/hy3)

Model card as data

GET /api/public/models/tencent/hy3Open
Machine-readable:/llms.txt/llms-full.txt