LLM Pricing
PricingLeaderboardsToolsProvidersReleasesGuides

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Hy3

tencent·tencent/hy3·GA·Open weights·Hy series
Tencent Hy reasoning model for coding, instruction following, and agent tasks

Hy3 by tencent is offered by 22 providers on this page. Public prices are shown for 19 of them. Its official list price is $0 per 1M tokens. The lowest paid channel is NanoGPT at $0.066 / $0.26 per 1M, about 2.7× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. It also has 1 free ($0) channel and 2 covered by a paid subscription; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.

Artificial Analysis rates it 25.3 on the Intelligence Index, with 58.8 for coding and 24.1 for agentic tasks. Against its lowest blended price of $0.115 per 1M, that is roughly 221 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 88 tokens per second, with 3.07s to the first token. Running one task of the Artificial Analysis suite costs about $0.0718, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 256,000 tokens at the reference host, but hosts report different limits, from 202,752 to 262,144, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use.

Specs & pricing

Input / output per 1M tokens
Official price·Tencent TokenHub
$0 / $0
Blended $0 · Cache read $0
Lowest paid·NanoGPTGateway
$0.066 / $0.26
Blended $0.11 · 2.7× spread
1 more $0 channels
Context
256,000
Output limit
128,000
Knowledge cutoff
—
Released / updated
2026-07-06 / 2026-07-06
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ TemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
Text
Weights
Hugging Face

Quality & performanceArtificial Analysis · Intelligence Index v4.3

Intelligence25.3
Value221
Coding58.8
Agentic24.1
Output speed88 tok/s
TTFT3.07 s
Cost per task$0.072
Value formulaIQ 25.3 ÷ min blended $0.115 = 221

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 22 providers19 with public prices · 1 free · 2 subscription-covered

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
Tencent TokenHubOfficialFirst-partySubscription$0$0256,000128,000
KenariGatewayFree——256,000128,000
Tencent Token PlanGatewaySubscription$0$0256,000128,000
NanoGPTGateway$0.066$0.26$0.029—262,144 ⚠128,000
Deep Infra
tencent/Hy3
Cloud$0.13$0.53$0.033—262,144 ⚠128,000
Kilo GatewayGateway$0.13$0.53$0.033—262,144 ⚠128,000
Eden AI
deepinfra/tencent/Hy3
Gateway$0.13$0.53$0.033—262,144 ⚠128,000
SiliconFlow
tencent/Hy3
Cloud$0.13$0.53$0.033—262,144 ⚠262,144
OpenRouterGateway$0.13$0.53$0.033—262,144 ⚠128,000
DevPass (LLM Gateway)Gateway$0.13$0.53$0.033—262,144 ⚠128,000

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

Toggle (on / off)effort = noneeffort = higheffort = lowbudget_tokens ≥ 128 ≤ 32,768effort = minimaleffort = mediumeffort = xhigheffort = max

Interleaved thinking (reasoning between tool calls) is declared by 2 of 22 providers.

1 / 22 providers expose no reasoning control (reasoning_options: []).

Your usage cost

1NanoGPT$21.76
2Deep Infra$40.86
3Kilo Gateway$40.86
4Eden AI$40.86
5SiliconFlow$40.92
6OpenRouter$40.92

3 more channels offer $0 (Tencent TokenHub, Kenari, Tencent Token Plan); free tiers usually have rate limits and no SLA, excluded from ranking.

The cheapest paid channel is the only channel.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$0.26$02026-08-052026-09-032026-08-05 · Input list $02026-08-06 · Input list $02026-08-07 · Input list $02026-08-08 · Input list $02026-08-09 · Input list $02026-08-10 · Input list $02026-08-11 · Input list $02026-08-12 · Input list $02026-08-13 · Input list $0.0662026-08-28 · Input list $02026-09-03 · Input list $02026-08-05 · Output list $02026-08-06 · Output list $02026-08-07 · Output list $02026-08-08 · Output list $02026-08-09 · Output list $02026-08-10 · Output list $02026-08-11 · Output list $02026-08-12 · Output list $02026-08-13 · Output list $0.262026-08-28 · Output list $02026-09-03 · Output list $02026-08-05 · Min blended $0.232026-08-06 · Min blended $0.232026-08-07 · Min blended $0.232026-08-08 · Min blended $0.232026-08-09 · Min blended $0.232026-08-10 · Min blended $0.232026-08-11 · Min blended $0.232026-08-12 · Min blended $0.232026-08-13 · Min blended $0.112026-08-28 · Min blended $0.232026-09-03 · Min blended $0.11

Benchmark31 items

NameConditionsScoreMetricSource
SWE-Bench Verified—78resolvedSource ↗
SWE-Bench Multilingualharness: SWE-agent · variant: highest reasoning effort75.8scoreSource ↗
SWE-Bench Proharness: SWE-agent · variant: highest reasoning effort57.9scoreSource ↗
Terminal-Benchharness: Terminus 2 · variant: highest reasoning effort; 4h timeout; 500 episodes · v2.171.7scoreSource ↗
NL2Repoharness: Claude Code · variant: highest reasoning effort; 250 turns; 12000s timeout45.6scoreSource ↗
DeepSWEharness: mini-swe-agent · variant: highest reasoning effort; 2h timeout28scoreSource ↗
BrowseCompharness: Tencent internal search harness · variant: highest reasoning effort84.2scoreSource ↗
WideSearchharness: Tencent internal search harness · variant: highest reasoning effort76.4scoreSource ↗
DeepSearchQAharness: Tencent internal search harness · variant: highest reasoning effort91scoreSource ↗
MCP Atlasharness: Scale AI · variant: highest reasoning effort; April 2026; 100 tool calls · dataset: 500 public tasks79.1scoreSource ↗
Toolathlonvariant: highest reasoning effort48.5scoreSource ↗
APEX-Agentsvariant: highest reasoning effort25.6pass@1Source ↗
ClawEvalharness: Tencent internal harness · variant: highest reasoning effort · dataset: 105 queries · v2026032568.5pass@3Source ↗
WildClawBenchharness: OpenClaw · variant: highest reasoning effort · dataset: 35 text-only tasks53.6scoreSource ↗
SkillsBenchharness: Claude Code · variant: highest reasoning effort · dataset: 79 text-only tasks55.3average over 3 runsSource ↗
Humanity's Last Examvariant: highest reasoning effort; with tools · dataset: text-only53.2scoreSource ↗
Humanity's Last Examvariant: highest reasoning effort; without tools · dataset: text-only37scoreSource ↗
GPQA Diamondvariant: highest reasoning effort90.4scoreSource ↗
FrontierSciencevariant: highest reasoning effort; research21.3scoreSource ↗
FrontierSciencevariant: highest reasoning effort; olympiad74.8scoreSource ↗
USAMOvariant: highest reasoning effort · v202672scoreSource ↗
MathArena Apexvariant: highest reasoning effort38.7scoreSource ↗
ArxivMathvariant: highest reasoning effort52.2scoreSource ↗
HorizonMathvariant: highest reasoning effort7.1pass@12Source ↗
PHYBenchvariant: highest reasoning effort77.4scoreSource ↗
CMT Benchmarkvariant: highest reasoning effort37.8scoreSource ↗
IMOAnswerBenchvariant: highest reasoning effort90scoreSource ↗
SuperChemvariant: highest reasoning effort54.9scoreSource ↗
CL-benchvariant: highest reasoning effort23.8scoreSource ↗
CL-bench-lifevariant: highest reasoning effort17scoreSource ↗
AA-LCRvariant: highest reasoning effort73.4scoreSource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Artificial Analysis evaluations6 items

GPQA Diamond89.7%
Humanity's Last Exam33.5%
SciCode48.6%
Terminal-Bench 2.164.4%
τ³-Bench Banking22.9%
AA-LCR79.0%

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.

Related models

Hy4 previewsame series$0.83 / $2.50Hy-MT2-1.8Bsame series$0.044 / $0.18Hy-MT2-30B-A3Bsame series$0.074 / $0.30
Data partly from models.dev (MIT) · Tencent TokenHub official docs ↗