# Claude Haiku 4.5 vs Hy4 preview

> Hy4 preview is the stronger model overall, scoring 44.6 to 39.3 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-haiku-4-5-vs-hy4-preview
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 0 categories and Hy4 preview in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Hy4 preview leads 30.6 to 15.1.
- The biggest single-benchmark swing is NYT Connections (extended): 14.3% for Claude Haiku 4.5 and 68.2% for Hy4 preview.
- Hy4 preview is cheaper at $0.83 / $2.50 per million input/output tokens, against $1 / $5 for Claude Haiku 4.5.
- Hy4 preview accepts more context: 1.05M tokens versus 200K.
- Hy4 preview has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| Provider | Anthropic | Tencent |
| Noometry Index | 39.3 | 44.6 |
| Rank | 155 | 73 |
| Context | 200K | 1.05M |
| Input $/M | $1 | $0.83 |
| Output $/M | $5 | $2.50 |
| Weights | Proprietary | Open |

## Coding

- Claude Haiku 4.5: 43.9 (#71)
- Hy4 preview: 51.4 (#36)

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| LMArena WebDev | 1330 | 1633 |
| SWE-bench Verified (bash only) | 66.6% | — |
| SWE-bench Multilingual | 64.7% | — |
| SciCode | 43.3% | — |
| WeirdML | 45.4% | — |
| LMArena Coding | 1454 | — |
| ALE-Bench | 653.48 | — |

## Agentic & Tool Use

- Claude Haiku 4.5: 33.5 (#50)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| Terminal-Bench | 35.5% | — |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| Vending-Bench 2 | 458.89 | — |

## Reasoning

- Claude Haiku 4.5: 15.1 (#310)
- Hy4 preview: 30.6 (#75)

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| NYT Connections (extended) | 14.3% | 68.2% |
| ARC-AGI-2 | 4% | — |
| ARC-AGI-1 | 47.7% | — |
| CritPt | 0% | — |
| Chess Puzzles | 8% | — |
| LMArena Hard Prompts | 1420 | — |
| DTBench | 73.6% | — |
| LMCA | 30.9% | — |
| Epoch Capabilities Index | 142.42 | — |
| ForecastBench | 61.4 | — |

## Math

- Claude Haiku 4.5: 43.0 (#69)
- Hy4 preview: 54.5 (#37)

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | — |
| ProofBench | — | 75% |
| Omni-MATH | 56.1% | — |
| LMArena Math | 1395 | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- Claude Haiku 4.5: 37.6 (#146)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| GPQA Diamond | 71.2% | — |
| SimpleQA Verified | 13.2% | — |
| MMLU-Pro | 77.7% | — |
| Vectara Hallucination Rate | 9.8% | — |
| GPQA (HELM) | 60.5% | — |
| LMArena Expert | 1443 | — |

## Multimodal

- Claude Haiku 4.5: 27.0 (#111)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |

## Multilingual

- Claude Haiku 4.5: 50.1 (#124)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| LMArena Non-English | 1376 | — |
| LMArena Chinese | 1415 | — |
| LMArena French | 1408 | — |
| LMArena German | 1377 | — |
| LMArena Japanese | 1339 | — |
| LMArena Korean | 1346 | — |
| LMArena Russian | 1380 | — |
| LMArena Spanish | 1420 | — |

## Instruction Following

- Claude Haiku 4.5: 71.6 (#144)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| IFEval | 80.1% | — |
| LMArena Instruction Following | 1414 | — |

## Long Context

- Claude Haiku 4.5: 43.6 (#86)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| LMArena Longer Query | 1427 | — |

## Writing & Preference

- Claude Haiku 4.5: 58.1 (#119)
- Hy4 preview: —

| Benchmark | Claude Haiku 4.5 | Hy4 preview |
|---|---|---|
| LMArena Text | 1396 | — |
| LMArena Creative Writing | 1371 | — |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
| LMArena Multi-Turn | 1409 | — |

## FAQ

### Is Claude Haiku 4.5 better than Hy4 preview?

Hy4 preview is the stronger model overall, scoring 44.6 to 39.3 on the Noometry Index.

### Which is cheaper, Claude Haiku 4.5 or Hy4 preview?

Hy4 preview is cheaper. It lists at $0.83 per million input tokens and $2.50 per million output tokens; Claude Haiku 4.5 lists at $1 and $5.

### Is Claude Haiku 4.5 or Hy4 preview better for coding?

Hy4 preview scores higher on coding benchmarks: 51.4 versus 43.9 in the Noometry coding category.

### Which has the bigger context window?

Hy4 preview does, with 1.05M tokens against 200K.

### How many benchmarks do Claude Haiku 4.5 and Hy4 preview share?

2 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and Hy4 preview has 3.