# Hy4 preview vs Qwen3 32B

> Hy4 preview is the stronger model overall, scoring 44.6 to 38.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/hy4-preview-vs-qwen3-32b
- Last updated: 2026-10-10
- Shared benchmarks: 0

## Summary

- The widest gap is in math, where Hy4 preview leads 54.5 to 37.9.
- Both cost about the same: $0.83 input and $2.50 output per million tokens.
- Hy4 preview accepts more context: 1.05M tokens versus 131K.

## Snapshot

| | Hy4 preview | Qwen3 32B |
|---|---|---|
| Provider | Tencent | Alibaba (Qwen) |
| Noometry Index | 44.6 | 38.8 |
| Rank | 73 | 166 |
| Context | 1.05M | 131K |
| Input $/M | $0.83 | $0.70 |
| Output $/M | $2.50 | $2.80 |
| Weights | Open | Open |

## Coding

- Hy4 preview: 51.4 (#36)
- Qwen3 32B: 37.3 (#184)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| Aider Polyglot | — | 40% |
| LMArena WebDev | 1633 | — |
| SciCode | — | 35.4% |
| LMArena Coding | — | 1358 |

## Agentic & Tool Use

- Hy4 preview: —
- Qwen3 32B: 32.3 (#58)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |

## Reasoning

- Hy4 preview: 30.6 (#75)
- Qwen3 32B: 19.6 (#233)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | — | 54.9% |
| NYT Connections (extended) | 68.2% | — |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| LMArena Hard Prompts | — | 1335 |
| DTBench | — | 67.5% |
| LMCA | — | 17.3% |
| Epoch Capabilities Index | — | 138.51 |

## Math

- Hy4 preview: 54.5 (#37)
- Qwen3 32B: 37.9 (#89)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 66.9% |
| ProofBench | 75% | — |
| LMArena Math | — | 1399 |

## Knowledge

- Hy4 preview: —
- Qwen3 32B: 40.0 (#112)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| GPQA Diamond | — | 65.7% |
| Vectara Hallucination Rate | — | 5.9% |
| LMArena Expert | — | 1362 |

## Multilingual

- Hy4 preview: —
- Qwen3 32B: 45.8 (#162)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| LMArena Non-English | — | 1317 |
| LMArena Chinese | — | 1357 |
| LMArena German | — | 1341 |
| LMArena Russian | — | 1311 |

## Instruction Following

- Hy4 preview: —
- Qwen3 32B: 69.1 (#174)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | — | 1305 |

## Long Context

- Hy4 preview: —
- Qwen3 32B: 43.7 (#83)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | — | 74.2% |
| LMArena Longer Query | — | 1327 |

## Writing & Preference

- Hy4 preview: —
- Qwen3 32B: 53.1 (#158)

| Benchmark | Hy4 preview | Qwen3 32B |
|---|---|---|
| LMArena Text | — | 1340 |
| LMArena Creative Writing | — | 1297 |
| LMArena Multi-Turn | — | 1331 |

## FAQ

### Is Hy4 preview better than Qwen3 32B?

Hy4 preview is the stronger model overall, scoring 44.6 to 38.8 on the Noometry Index.

### Which is cheaper, Hy4 preview or Qwen3 32B?

Qwen3 32B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; Hy4 preview lists at $0.83 and $2.50.

### Is Hy4 preview or Qwen3 32B better for coding?

Hy4 preview scores higher on coding benchmarks: 51.4 versus 37.3 in the Noometry coding category.

### Which has the bigger context window?

Hy4 preview does, with 1.05M tokens against 131K.

### How many benchmarks do Hy4 preview and Qwen3 32B share?

0 benchmarks have published results for both models. Hy4 preview has 3 scored results on Noometry and Qwen3 32B has 26.