Hunyuan 3

Hunyuan 3 is Tencent's open-weight mixture-of-experts model, released under Apache 2.0 for software development, office productivity, and financial modeling.

input · usd / 1m tok

$0.132

output · usd / 1m tok

$0.528

context window

262.14ktokens

modalities · in → out

takes text in, gives text out

pricing

base api ratesusd per 1m tokens

input
$0.132
output
$0.528
cache read
$0.033

providers

Different companies host Hunyuan 3. Impossibl routes your request intelligently and handles failover automatically when a provider is down.

  • Tencent
  • AtlasCloud

specifications

takes text in, gives text outtencent/hy3

  • creatorTencent
  • model typeText generation
  • context window262,144 tokens
  • maximum outputnot published
  • input

    takes text in, gives text out

  • outputtext
  • tool callingnot verified
  • releasedjul 6, 2026
  • endpoint/v1/chat/completions

unknown means the public catalog does not establish that fact. a model's documented capabilities can differ from what a particular api supports. see the model documentation for integration details.

try it · your prompt, this model, live

# responses API
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["IMPOSSIBL_API_KEY"],
    base_url="https://api.impossibl.com/v1",
)

resp = client.responses.create(
    model="tencent/hy3",
    input="",
)
print(resp.output_text)

edit the prompt, then run

benchmarks

intelligence index

25.3#58 of 108 models on the gateway

  1. coding index58.8#35 of 64
  2. agentic index24.1#31 of 53
  3. omniscience-18.5#65 of 84
  4. gpqa89.7%#28 of 80
  5. humanity’s last exam33.5%#38 of 87
  1. GPQA89.7%
  2. AA-LCR79.0%
  3. SciCode48.6%
  4. Humanity’s Last Exam33.5%
  5. Tau Banking22.9%
  6. AutomationBench Partial18.4%
  7. GDP.pdf All-pass11.2%
  8. CritPt4.9%
  • Intelligence Index25.30
  • Omniscience Index-18.47
  • GDPval Elo1066.97
  • legal index (API)26.5
  • coding index (API)58.8
  • agentic index (API)24.1
  • economics index (API)35.2
  • engineering index (API)26.7
  • intelligence index (API)25.3
  • strategy and ops index (API)23.9
  • finance and accounting index (API)27.3

view source ↗all source data and breakdowns (json) ↗

each configuration is listed separately; these evaluations describe the tested model configuration, and performance and costs can differ on impossibl. unmeasured scores are omitted and api values keep their source units. last checked 2026-10-14.

compare Hunyuan 3 with other models

frontier labs and open weights, best-ranked firstusd / 1m tok · in / out

impossibl models --all →

frequently asked questions

What is Hunyuan 3?

Hunyuan 3 is Tencent's open-weight mixture-of-experts model, released under Apache 2.0 for software development, office productivity, and financial modeling. Over Hy2 it improves complex reasoning, instruction following, code generation, and agent capabilities, and exposes configurable reasoning effort.

How much does Hunyuan 3 cost?

Hunyuan 3 costs $0.132 per million input tokens and $0.528 per million output tokens through the Impossibl API. Cache reads cost $0.033 per million tokens. All rates are in USD.

view the pricing table ↑

What is the context length of Hunyuan 3?

Hunyuan 3 has a 262,144-token context window through the Impossibl API.

What inputs and outputs does Hunyuan 3 support?

Hunyuan 3 accepts text as input. Hunyuan 3 returns text.

How do I use Hunyuan 3 through an API?

Hunyuan 3 is available through the Impossibl API. Use tencent/hy3 as the model identifier with /v1/chat/completions. The quickstart explains how to create an API key and send your first request.

follow the api quickstart →

When was Hunyuan 3 released?

Tencent released Hunyuan 3 on July 6, 2026.

this page describes Hunyuan 3 as listed by the impossibl public api. prices and availability come from that catalog, refreshed here every five minutes. this is an impossibl offering, not a comparison of independent hosting-provider quotes. model descriptions and release dates, where available, use the creator sources linked under the answers above.

view the source catalog ↗model as markdown ↗report a correction →