priorsBuilder

Tencent · released 2026-07-06

Hy3

Measured on 32 benchmarks across 5 reasoning efforts.

General frontier

32.8

#74 of 136 models · best at none

includes estimates: 5% of the mix is measured for this model

Reasoning efforts in your mix

none32.8#202 of 313
default32.3#203 of 313
xhigh28.6#230 of 313
default25.3#247 of 313
none15.2#297 of 313

Your mix’s benchmarks

3 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #131default of 58733.5%
  2. #362none of 5877.0%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #232default of 526-18.5
  2. #289none of 526-35.0

Leader: Opus 5.5 · 46.4

reasoning

BullshitBench v2

196 results
  1. #156none of 19616.0%
  2. #179xhigh of 1968.0%

Leader: Opus 4.8 · 95.0%

All 32 benchmarks

show 29 more

agentic

AA-Briefcase v1.1

224 results
  1. #126default of 224947

Leader: Sonnet 5.5 · 1823

agentic

Agent Arena

50 results
  1. #40xhigh of 50-5.1%

Leader: Opus 5.5 · 15.5%

agentic

AutomationBench-AA

219 results
  1. #135default of 21918.4%

Leader: Gemini 4 · 77.5%

agentic

τ²-Bench Telecom

393 results
  1. #49default of 39392.7%
  2. #168none of 39367.5%

Leader: GLM 5.2 · 99.1%

agentic

τ³-Banking

206 results
  1. #89default of 20622.9%

Leader: Qwen 3.8 Max · 51.3%

coding

Arena WebDev

122 results
  1. #53xhigh of 1221508

Leader: Opus 5.5 · 1813

coding

SciCode

226 results
  1. #119default of 22648.6%

Leader: Opus 5.5 · 66.9%

coding

Terminal-Bench Hard-AA

387 results
  1. #88default of 38734.1%
  2. #104none of 38731.8%

Leader: GPT-5.6 Sol · 65.9%

coding

Terminal-Bench v2.1-AA

238 results
  1. #85default of 23864.4%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #137default of 2160.5%

Leader: Sonnet 5.5 · 63.6%

composite

AA Intelligence Index v4.3.2

610 results
  1. #157default of 61025.3
  2. #257none of 61017.0

Leader: Opus 5.5 · 57.6

composite

Arena Text

220 results
  1. #63xhigh of 2201454

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #76xhigh of 2201451

Leader: Gemini 4 · 1553

cost

AA Intelligence Index Cost per Task

177 results
  1. #32default of 177$0.07

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #35default of 177$150

Leader: GPT-6 Luna · $10.63

economics

GDP.pdf

210 results
  1. #120default of 21011.2%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #144default of 2881067

Leader: Opus 5.5 · 1866

instruction

IFBench

399 results
  1. #115default of 39963.1%
  2. #184none of 39948.0%

Leader: Grok 4.3 · 83.3%

knowledge

AA-Omniscience Accuracy

526 results
  1. #165default of 52632.0%
  2. #253none of 52623.2%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #230default of 52625.9%
  2. #244none of 52624.3%

Leader: MiniCPM5-1B (Non-reasoning) · 99.1%

knowledge

GPQA Diamond

539 results
  1. #70default of 53989.7%
  2. #266none of 53973.2%

Leader: GPT-6 Astra · 96.3%

long-context

AA-LCR v1.1

523 results
  1. #106default of 52379.0%
  2. #334none of 52342.0%

Leader: Kimi K3 · 88.7%

reasoning

CritPt

529 results
  1. #153default of 5294.9%
  2. #274none of 5290.3%

Leader: Muse Wonder Full (Max) · 32.9%

reliability

Tool-call error rate

171 results
  1. #77xhigh of 1711.1%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #82xhigh of 18199.6%

Leader: GPT-5 Pro · 100.0%

speed

Latency (Time to First Token)

351 results
  1. #197default of 3512.9s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #196default of 35184.3

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #132xhigh of 1772.4s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #92xhigh of 17661.0

Leader: Muse Spark 1.1 · 204.3