Singularity Logs

THE MODEL, IN DETAIL / tencent

Hy3

tencent/hy3Open weightsReasoning

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:...

OpenRouter
Source-linked model intelligenceCatalog added 6 Jul 2026Specifications observed 17 Oct 2026

Overall agents

#44

Arena agent score

Context window

262.1K

128K max output tokens

Input price

$0.132

USD / million input tokens

Output price

$0.528

USD / million output tokens

Inference providers

5

5 provider routes observed

Measured output speed

—

No model-level measurement

01 / PERFORMANCE

A clearer view of capability.

Read the score, understand the context, and explore where this model sits among its peers.

Performance fingerprint

Individual published accuracy scores · not a composite index

At least three published accuracy scores are needed for this view.

In good company

Overall agents · Arena · official dataset

17.2
Anthropic
14.8
OpenAI
14.1
Anthropic
13.7
Google
13.7
OpenAI
12.5
Anthropic
11.2
OpenAI
9.2
Anthropic
8.4
Anthropic
7.4
OpenAI

54 models in the source cohort. Showing selected nearby models. Evaluation settings can differ.

Evidence, by task.

Coverage is not a capability score.

0published results
0 evaluation sources
ReasoningAwaiting evidence
CodingAwaiting evidence
MathematicsAwaiting evidence
VisionAwaiting evidence
AgenticAwaiting evidence
Long contextAwaiting evidence

Documented features, task evaluations, and preference rankings describe different things. Missing coverage does not imply a model cannot perform a task.

Explore our methodology

Where people rank it

Published rankings for this exact model or configuration. Each source keeps its original metric.

Overall agentsArena agent score · 14 Oct 2026#44-6.1%Confidence interval: -7.7–-4.6Agents for workTask-outcome score · 14 Oct 2026#48-10.5%Confidence interval: -13.4–-7.6

02 / BENCHMARK EXPLORER

The numbers behind the model.

Original results, exact versions, and a direct path to the evidence behind every score.

0 published results

Percentage-based evaluations will appear here when published.

0 percentage-based evaluations. Each benchmark measures a different task; compare results within the same evaluation and protocol.

Design Arena

Source-reported Elo across individual design categories

3d1,176
models arena · Rank #57 · 44% win rate
Codecategories1,178
models arena · Rank #67 · 41.2% win rate
Dataviz1,125
models arena · Rank #91 · 36.1% win rate
Gamedev1,139
models arena · Rank #79 · 38.6% win rate
Uicomponent1,163
models arena · Rank #72 · 40.2% win rate
Website1,183
models arena · Rank #71 · 41.3% win rate

Catalog observation: 17 Oct 2026. Category ratings have separate comparison pools.

03 / COST & VALUE

Know what your workload costs.

Explore token prices, see where the spend goes, and model a workload of your own.

Input$0.132per million tokens
Output$0.528per million tokens
Cache read$0.033per million tokens
Cache write—per million tokens
Additional published rates
Additional rates supplied by the model catalog
Billing fieldPublished rateBilling basis
Overrides[{"prompt":"0.000000132","utc_end":1600,"utc_start":0,"completion":"0.000000528","input_cache_read":"0.000000033"},{"prompt":"0.0000000825","utc_end":0,"utc_start":1600,"completion":"0.00000033","input_cache_read":"0.000000020625"}]Source-supplied USD rate; see provider billing terms

Workload cost calculator

Change the numbers. See the trade-offs.

Token estimate only. Excludes cache writes, tools, and other provider fees. Actual token usage and provider limits vary.

Your estimated spend

1,000 requests · 11M tokens

$1.848estimated total
Uncached input$1.32
Cached input$0
Output$0.528
Per request$0.0018
Without cache discount$1.848
Estimated cache savings$0

04 / INFERENCE NETWORK

One model. Different ways to run it.

Compare inference routes by price, context, quantization, and reported availability.

5 providers · 5 routes
Provider routes for Hy3, observed 17 Oct 2026
Provider / routeInput / MOutput / MCache / MContextOutput limitSpeedLatencyAvailability · 1dRates
TTencentFP8$0.132$0.528$0.033262.1K128K——100%
NNovitaQuantization not disclosed$0.14$0.58$0.035262.1K235.9K——100%
GGMICloudBF16$0.14$0.58$0.035262.1K235.9K——99.52%
PPhalaQuantization not disclosed$0.15$0.64$0.04262.1K235.9K——100%
AAtlasCloudFP8$0.2$0.8$0.05262.1K131.1K——100%

Snapshot observed 17 Oct 2026. Speed and latency are the reported 30-minute median when available. This is a dated snapshot, not live monitoring.

Provider price comparison

USD per million tokens · input and output rates

InputOutput
Tencentfp8
Input
$0.132
Output
$0.528
Novita
Input
$0.14
Output
$0.58
GMICloudbf16
Input
$0.14
Output
$0.58
Phala
Input
$0.15
Output
$0.64
AtlasCloudfp8
Input
$0.2
Output
$0.8

Reported availability

Provider availability over the selected observation window

Tencent · FP8100%
Novita100%
GMICloud · BF1699.52%
Phala100%
AtlasCloud · FP8100%

Output speed

Reported 30-minute p50 · tokens per second

No provider throughput measurements were returned in this snapshot.

Response latency

Reported 30-minute p50 · seconds

No provider latency measurements were returned in this snapshot.

Context by provider

Actual route limits can differ from the catalog maximum

Tencent · FP8262.1K
Novita262.1K
GMICloud · BF16262.1K
Phala262.1K
AtlasCloud · FP8262.1K

05 / CAPABILITIES & MODALITIES

What goes in. What comes out.

A map of documented interfaces and API features, separate from benchmark performance.

INPUT MODALITIES

Text
Hy3Reasoning supported

OUTPUT MODALITIES

Text

Feature support

Listed catalog API parameters and documented pricing features

ReasoningSupported
Tool callingSupported
Tool choiceSupported
Structured outputsSupported
Response formatSupported
Reasoning detailsSupported
Seed controlSupported
Stop sequencesSupported
Prompt cachingSupported
TemperatureSupported

Room to work

Context capacity and completion limits

262.1Ktotal context tokens
Context window
128Kmax output

Output tokens share the total context budget. The maximum completion length is a separate limit, and individual provider routes may have smaller limits.

Catalog context
262,144 tokens
Maximum output
128,000 tokens
Top-provider context
262,144 tokens

Features vary by route

Check the provider before depending on a specific API feature.

API feature support by provider
Provider routeReasoningToolsStructured outputResponse formatImplicit caching
Tencent · FP8SupportedSupportedSupportedSupportedSupported
NovitaSupportedSupportedSupportedNot listedSupported
GMICloud · BF16SupportedSupportedNot listedSupportedSupported
PhalaSupportedSupportedSupportedNot listedSupported
AtlasCloud · FP8SupportedSupportedSupportedNot listedSupported

06 / UNDER THE HOOD

The complete specification sheet.

Exact identities, model architecture, limits, and the defaults that shape a request.

Identity & architecture

Developer
tencent
Exact identity
Hy3
Subject
Agent System
Access
Open weights
License
Apache 2.0
Parameters
Not disclosed
Tokenizer
Other
Instruction format
Not disclosed

Reasoning & controls

Reasoning
Supported
Mandatory reasoning
No
Reasoning by default
Enabled
Published configuration
Not disclosed
Supported reasoning efforts
high, low, none
Default reasoning effort
high
API parameters
19
Provider moderation
No
Quantization variants
fp8, bf16

Dates & limits

Catalog added
6 Jul 2026
Knowledge cutoff
Not disclosed
Catalog expiration
Not documented
Specifications observed
17 Oct 2026
Providers observed
17 Oct 2026
Request limits
Provider-specific / not supplied
Canonical catalog ID
tencent/hy3-20260706

API parameters & defaults

Availability is catalog-wide. Defaults can vary across providers.

frequency_penaltySupported · default not supplied
include_reasoningSupported · default not supplied
logit_biasSupported · default not supplied
max_completion_tokensSupported · default not supplied
max_tokensSupported · default not supplied
min_pSupported · default not supplied
presence_penaltySupported · default not supplied
reasoningSupported · default not supplied
reasoning_effortSupported · default not supplied
repetition_penaltySupported · default not supplied
response_formatSupported · default not supplied
seedSupported · default not supplied
stopSupported · default not supplied
structured_outputsSupported · default not supplied
temperatureDefault: 0.9
tool_choiceSupported · default not supplied
toolsSupported · default not supplied
top_kSupported · default not supplied
top_pDefault: 1

07 / BUILD WITH THIS MODEL

From exploration to your first request.

Use the documented catalog identifier and keep your API key in an environment variable.

Make an API request

const response = await fetch(
  "https://openrouter.ai/api/v1/chat/completions",
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.OPENROUTER_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      "model": "tencent/hy3",
      "messages": [
        {
          "role": "user",
          "content": "Explain the trade-offs in this design."
        }
      ]
    }),
  },
);
if (!response.ok) throw new Error(`HTTP ${response.status}`);
const data = await response.json();
console.log(data.choices[0].message.content);

08 / PROVENANCE & HISTORY

Every number has a story.

Follow the source, check the observation date, and distinguish published results from model specifications.

The observation trail

Recorded events, in chronological order

Added to the model catalog

Catalog creation date, as recorded by OpenRouter.

Model specifications observed

Catalog pricing, limits, and supported parameters captured.

Provider routes observed

5 inference routes in this snapshot.

Observation dates describe when data was captured. They do not represent newly run evaluations or continuous performance monitoring.

Published observation ledger

Historical evidence, with source and version preserved.

No published benchmark observations for this selection. Published preference rankings appear separately above.

09 / A FEW MORE DETAILS

Questions worth asking.

Quick answers, drawn from this model’s connected specifications.

How much does Hy3 cost?

Catalog starting rates are $0.132 per million input tokens and $0.528 per million output tokens. Cache-read pricing is $0.033 per million tokens. Compare provider routes and service tiers above for their specific prices.

How much context can it handle?

262,144 tokens in the catalog context window. The catalog lists a maximum of 128,000 completion tokens. Provider limits can differ.

Does it support tools and structured output?

Tool calling: listed as supported. Structured outputs: listed as supported. See the provider feature matrix for route-specific differences.

How should I interpret these benchmark results?

Read the benchmark name, exact version, metric, source, and model configuration together. Different evaluation settings can change the result. Preference ratings and source-defined composite indices remain separate from task accuracy.

Are these prices and provider measurements live?

This page uses dated source snapshots. Catalog data was observed on 17 Oct 2026. Provider routes were observed on 17 Oct 2026. Missing measurements are not estimated or treated as zero.

KEEP EXPLORING

More from tencent.

Explore related identities from the same developer, or take them into a side-by-side comparison.

THE RIGHT MODEL STARTS WITH THE RIGHT CONTEXT.

Put the details side by side.

Explore the trade-offs in a direct model comparison.

Compare Hy3
Source-backed evidence. Different tasks. A clearer picture.
AboutMethodologyBenchmark dashboard