# Claude 3 Haiku vs Gemma 3 4B

> Gemma 3 4B is the stronger model overall, scoring 27.9 to 25.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-haiku-vs-gemma-3-4b
- Last updated: 2026-10-10
- Shared benchmarks: 19

## Summary

- They share 19 benchmarks with published results for both. Claude 3 Haiku scores higher in 2 categories and Gemma 3 4B in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 3 4B leads 42.2 to 29.8.
- The biggest single-benchmark swing is GPQA Diamond: 36.3% for Claude 3 Haiku and 23.2% for Gemma 3 4B.
- Gemma 3 4B has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| Provider | Anthropic | Google |
| Noometry Index | 25.7 | 27.9 |
| Rank | 333 | 320 |
| Context | — | 131K |
| Input $/M | — | $0.04 |
| Output $/M | — | $0.08 |
| Weights | Proprietary | Open |

## Coding

- Claude 3 Haiku: 26.1 (#320)
- Gemma 3 4B: 35.3 (#214)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1199 | 1230 |
| WeirdML | 9.8% | — |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |

## Agentic & Tool Use

- Claude 3 Haiku: —
- Gemma 3 4B: 20.7 (#134)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 19.6% |

## Reasoning

- Claude 3 Haiku: 15.7 (#303)
- Gemma 3 4B: 12.7 (#330)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 25.2% |
| LMArena Hard Prompts | 1174 | 1253 |
| DTBench | 50.1% | 50.9% |
| LMCA | 8.8% | 2.8% |
| Epoch Capabilities Index | 118.3 | 116.02 |
| Chess Puzzles | — | 0% |
| ForecastBench | 53.2 | — |
| WinoGrande | 74.2% | — |

## Math

- Claude 3 Haiku: 9.2 (#313)
- Gemma 3 4B: 15.8 (#286)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 7.5% |
| LMArena Math | 1188 | 1239 |
| MATH Level 5 | 14.9% | — |

## Knowledge

- Claude 3 Haiku: 17.3 (#280)
- Gemma 3 4B: 11.8 (#294)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| GPQA Diamond | 36.3% | 23.2% |
| LMArena Expert | 1148 | 1223 |
| Confabulations | 34.2% | — |
| Vectara Hallucination Rate | — | 6.4% |
| MMLU | 73.8% | — |

## Multimodal

- Claude 3 Haiku: 23.1 (#123)
- Gemma 3 4B: —

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |

## Multilingual

- Claude 3 Haiku: 36.1 (#238)
- Gemma 3 4B: 42.7 (#189)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1178 | 1273 |
| LMArena German | 1174 | 1281 |
| LMArena Russian | 1204 | 1294 |
| LMArena Chinese | 1155 | — |
| LMArena French | 1194 | — |
| LMArena Japanese | 1102 | — |
| LMArena Korean | 1108 | — |
| LMArena Spanish | 1165 | — |

## Instruction Following

- Claude 3 Haiku: 61.5 (#242)
- Gemma 3 4B: 65.4 (#219)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1174 | 1239 |

## Long Context

- Claude 3 Haiku: 36.1 (#232)
- Gemma 3 4B: 38.6 (#189)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1191 | 1274 |

## Writing & Preference

- Claude 3 Haiku: 29.8 (#286)
- Gemma 3 4B: 42.2 (#234)

| Benchmark | Claude 3 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1195 | 1291 |
| LMArena Creative Writing | 1157 | 1270 |
| EQ-Bench Creative Writing | 717 | 1068 |
| LMArena Multi-Turn | 1190 | 1255 |

## FAQ

### Is Claude 3 Haiku better than Gemma 3 4B?

Gemma 3 4B is the stronger model overall, scoring 27.9 to 25.7 on the Noometry Index.

### Is Claude 3 Haiku or Gemma 3 4B better for coding?

Gemma 3 4B scores higher on coding benchmarks: 35.3 versus 26.1 in the Noometry coding category.

### How many benchmarks do Claude 3 Haiku and Gemma 3 4B share?

19 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Gemma 3 4B has 22.