stealthmodels.

MODEL BENCHMARKS

Hy4 preview

Apache-2.0 MoE preview with 770B total parameters, 49B active and a million-token context. Sparse attention and multi-token prediction support long-input workloads. Very heavy in thinking on each instruction turn.

BENCHMARK BY BENCHMARK

Hy4 preview benchmark scores

StealthMark
43.9/100Overall rank 34/58
General Intelligence
41.3/100
Visual Intelligence
48.6/100

Hy4 previewOther modelsMedian

Reasoning efficiency

How many reasoning tokens a model needs for the quality it delivers.

Single-step reasoningHow much it reasons to understand a taskRank 51/56
Multi-step reasoningHow much it keeps reasoning as a task goes onRank 42/54

Cost efficiency

How much Performance you get for your money.

Rank 48/52

DATASHEET

Hy4 preview specs and pricing

A good fit forLong-document analysis and research on large GPU systems, with the freedom to inspect and adapt the weights.

Hy4 previewOther modelsMedian

PriceDeveloper API, input / output per 1M tokens
$0.834 / $2.501Cheaper than 25 of 46 others
Context windowTokens in one request
1,000,000Larger than 7 of 49 others
Output speedTokens per second, our measurement
64 tokens/sFaster than 3 of 51 others
Developer
Tencent Hy Team
Released
28 Aug 2026
Family
Hy4
License
Apache-2.0
Open model
Yes
Local model
No
Origin
China
Parameters
770B
Active parameters
49B
Max output
64,000 tokens
Input → output
Text → text
Architecture
MoE; 78-layer backbone with Gated DSA, 256 routed experts + 1 shared expert, top-8 routed experts active, plus native MTP layer

Self-host Hy4 preview

Hardware
Rack-scaleEstimated
Download size
814 GB (FP8)

Official Hugging Face weightsOfficial FP8 Hugging Face weights

All prices and providers

Tencent API; input 0.834 USD/1M; cached read 0.042/1M; output 2.501/1M; 10% lower from 16:00 to 24:00 UTC

  • TencentObserved 2026-10-23
    Input $0.834Output $2.501Cached read $0.042

    per 1M tokens

    Available

    tencent/hy4-preview
Deployment details
  • Official Hy4-preview-FP8 repository814 GB · FP8
Measurements and sources6

GAME BENCHMARKS

Watch Hy4 preview play

Foresight

Anticipating threats and planning ahead

53.1/100Rank 17/54
Survived 1 of 3
Foresight: complete playing field
Level 311 ptsTurn 16 / 20

Points per level

  1. Level 1Second Thoughts26 ptsBest 27Survived
  2. Level 2Pressure Pot6 ptsBest 13Out of lives
  3. Level 3Hide or Flight11 ptsBest 19Out of lives

Collect coins while avoiding ghosts. Ghosts move before the player each turn, and collisions cost a life. 3 levels in sequence, up to 66 turns in total. Earlier moves and outcomes carry forward, testing learning across levels.

Latch

Cooperative planning and learning from experience

35.6/100Rank 38/55
Cleared 3 of 5
Latch: complete playing field
Level 518.2 ptsTurn 19 / 40

Points per level

  1. Level 1Hold the door5.2 ptsBest 5.2Cleared
  2. Level 2Crossfeed6 ptsBest 10.2Finished early
  3. Level 3Relay vault18 ptsBest 19.2Cleared
  4. Level 4Passing bay15.8 ptsBest 18.2ClearedFinished early · 3 beacons left
  5. Level 5Second thoughts18.2 ptsBest 20.8Finished early

Coordinate two robots to push numbered cores onto matching sockets. Pressure plates hold doors open; moving either robot uses one turn. 5 levels in sequence, up to 152 turns in total. Earlier moves and outcomes carry forward, testing learning across levels.

Hy4 preview benchmark results ↗

StealthMark score

StealthMark measures how well a model’s abilities carry over to original, demanding tasks it could not have trained for: physical and spatial reasoning, planning across many steps, and precise visual work. How StealthMark works explains every benchmark, what it tests and why it is hard.

General Intelligence

The ability to reason, plan, solve problems and understand how things relate and behave. A high score means it thinks hard problems through, from careful analysis to plans that span many steps.

Visual Intelligence

The ability to understand visual intent and turn ideas, descriptions and plans into coherent, effective visuals. A high score means design ideas come out the way you meant them, in illustrations, icons, diagrams or interface visuals.

Overall score

Overall shows how well a model’s abilities carry over to original, demanding tasks, weighting every benchmark into one score. Estimated means the total includes an estimate for one missing scene.

Judging

Scores reflect how well each result meets the benchmark requirements. Game scores come from the recorded play: the points earned on each level and how each level ended. How StealthMark keeps scores honest.

Games

Each game is a sequence of levels, and earlier moves carry into later levels. A level that was not played is shown as not played and never counts as zero. Play every level of Latch or Foresight yourself to get a score calculated like the models’.

Reasoning efficiency

Good results with little thinking score high. On each task, a model’s result and reasoning tokens are compared with every other model and combined with the Balanced Integration Score (Liesefeld & Janczyk, 2019); 50 is average. Single step shows how much reasoning a model needs to understand and solve one problem, multi-step how efficiently it keeps working on a task by building on its earlier reasoning, as in agentic workflows. Not part of StealthMark.

Cost efficiency

How much StealthMark a model delivers for its money, in StealthMark points per dollar: its overall score divided by what one benchmark task costs it on average. The cost is what our runs were charged, or their tokens at the model’s list price; benchmarks a model has not run yet are estimated from its other costs. Like StealthMark, it is absolute and compares with no other model. Not part of StealthMark.

BENCHMARK RESULTS

Compare results

This link opens the same results in the same order.

StealthMark benchmark results

Reset benchmark preferences?

Return to the default view, clear comparison selections and restore the advertisement.