# Claude Haiku 5.5 vs DeepSeek V4 Flash

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 49.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-haiku-5-5-vs-deepseek-v4-flash
- Last updated: 2026-10-12
- Shared benchmarks: 8

## Summary

- They share 8 benchmarks with published results for both. Claude Haiku 5.5 scores higher in 2 categories and DeepSeek V4 Flash in 2 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 35.2.
- The biggest single-benchmark swing is NYT Connections (extended): 65.7% for Claude Haiku 5.5 and 89.6% for DeepSeek V4 Flash.
- Claude Haiku 5.5 is cheaper at $0.10 / $0.50 per million input/output tokens, against $0.15 / $0.60 for DeepSeek V4 Flash.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 49.5 | 53.6 |
| Rank | 52 | 35 |
| Context | 1M | 1M |
| Input $/M | $0.10 | $0.15 |
| Output $/M | $0.50 | $0.60 |
| Weights | Proprietary | Open |

## Coding

- Claude Haiku 5.5: 49.2 (#50)
- DeepSeek V4 Flash: 47.9 (#59)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| LMArena WebDev | 1587 | 1582 |
| FrontierCode | — | 18.8% |
| SciCode | — | 49.9% |
| WeirdML | — | 63% |
| LMArena Coding | — | 1457 |
| ALE-Bench | — | 1,306 |

## Reasoning

- Claude Haiku 5.5: 35.2 (#69)
- DeepSeek V4 Flash: 53.7 (#30)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| NYT Connections (extended) | 65.7% | 89.6% |
| Mystery Game Puzzles | 30% | 34% |
| ARC-AGI-2 | — | 61.4% |
| SimpleBench | — | 61.1% |
| Kagi LLM Benchmark | — | 52.2% |
| ARC-AGI-1 | — | 89% |
| CritPt | — | 16.6% |
| Chess Puzzles | — | 33% |
| LMArena Hard Prompts | — | 1444 |
| DTBench | — | 90.9% |
| LMCA | — | 41.7% |
| Epoch Capabilities Index | — | 154.49 |

## Math

- Claude Haiku 5.5: 73.6 (#18)
- DeepSeek V4 Flash: 60.3 (#37)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | 75.1% | 57.5% |
| FrontierMath Tier 4 | 46.3% | 24.4% |
| OTIS Mock AIME 2024-2025 | 98.9% | 94.4% |
| MathArena Final-Answer Competitions | — | 76.5% |
| ProofBench | — | 56% |
| LMArena Math | — | 1427 |

## Knowledge

- Claude Haiku 5.5: 50.8 (#70)
- DeepSeek V4 Flash: 55.4 (#48)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| GPQA Diamond | 89.6% | 91% |
| SimpleQA Verified | 23.8% | 33.6% |
| LMArena Expert | — | 1441 |

## Multimodal

- Claude Haiku 5.5: 43.7 (#21)
- DeepSeek V4 Flash: —

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| Furniture Assembly | 47.5% | — |

## Multilingual

- Claude Haiku 5.5: —
- DeepSeek V4 Flash: 53.0 (#72)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| LMArena Non-English | — | 1420 |
| LMArena Chinese | — | 1468 |
| LMArena French | — | 1439 |
| LMArena German | — | 1418 |
| LMArena Japanese | — | 1406 |
| LMArena Korean | — | 1384 |
| LMArena Russian | — | 1428 |
| LMArena Spanish | — | 1436 |

## Instruction Following

- Claude Haiku 5.5: —
- DeepSeek V4 Flash: 74.9 (#81)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| LMArena Instruction Following | — | 1421 |

## Long Context

- Claude Haiku 5.5: —
- DeepSeek V4 Flash: 43.8 (#85)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| LMArena Longer Query | — | 1434 |

## Writing & Preference

- Claude Haiku 5.5: —
- DeepSeek V4 Flash: 63.8 (#61)

| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash |
|---|---|---|
| LMArena Text | — | 1432 |
| LMArena Creative Writing | — | 1403 |
| EQ-Bench Creative Writing | — | 1559 |
| LMArena Multi-Turn | — | 1449 |

## FAQ

### Is Claude Haiku 5.5 better than DeepSeek V4 Flash?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 49.5 on the Noometry Index.

### Which is cheaper, Claude Haiku 5.5 or DeepSeek V4 Flash?

Claude Haiku 5.5 is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; DeepSeek V4 Flash lists at $0.15 and $0.60.

### Is Claude Haiku 5.5 or DeepSeek V4 Flash better for coding?

Claude Haiku 5.5 scores higher on coding benchmarks: 49.2 versus 47.9 in the Noometry coding category.

### Which has the bigger context window?

Both accept 1M tokens.

### How many benchmarks do Claude Haiku 5.5 and DeepSeek V4 Flash share?

8 benchmarks have published results for both models. Claude Haiku 5.5 has 9 scored results on Noometry and DeepSeek V4 Flash has 41.