Model comparison
Claude 3.5 Haiku vs Gemma 3 4B
Claude 3.5 Haiku is the stronger model overall, scoring 29.2 to 28.1 on the Noometry Index.
Last verified . 17 shared benchmarks.
GoogleSummary
- They share 17 benchmarks with published results for both. Claude 3.5 Haiku scores higher in 4 categories and Gemma 3 4B in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Claude 3.5 Haiku leads 28.0 to 20.9.
- The biggest single-benchmark swing is GPQA Diamond: 38.1% for Claude 3.5 Haiku and 23.2% for Gemma 3 4B.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Claude 3.5 Haiku | Gemma 3 4B | |
|---|---|---|
| Provider | Anthropic | |
| Noometry Index | 29.2 | 28.1 |
| Released | 2024-10-22 | 2025-03-12 |
| Weights | Proprietary | Open |
| Context window | — | 131K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.04 |
| Output $ / M tokens | — | $0.08 |
| Results tracked | 49 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 3 4B leads
Claude 3.5 Haiku: 32.9 (#265), Gemma 3 4B: 35.9 (#215)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1286 | 1230 |
| Aider Polyglot | 28% | — |
| SciCode | 27.4% | — |
| WeirdML | 30.7% | — |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 51.4% | — |
| BigCodeBench Complete | 59% | — |
| CadEval | 32% | — |
Agentic & Tool Use Claude 3.5 Haiku leads
Claude 3.5 Haiku: 28.0 (#95), Gemma 3 4B: 20.9 (#142)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 19.6% |
| BALROG | 19.3% | — |
Reasoning Claude 3.5 Haiku leads
Claude 3.5 Haiku: 17.7 (#290), Gemma 3 4B: 13.2 (#335)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1253 |
| DTBench | 56.7% | 50.9% |
| Epoch Capabilities Index | 127.15 | 116.02 |
| Kagi LLM Benchmark | — | 25.2% |
| CritPt | 0% | — |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 28.1% | — |
| LiveBench Data Analysis | 48.5% | — |
| LMCA | — | 2.8% |
| LiveBench | 43.5% | — |
Math Gemma 3 4B leads
Claude 3.5 Haiku: 14.7 (#300), Gemma 3 4B: 16.8 (#292)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 7.5% |
| LMArena Math | 1244 | 1239 |
| Omni-MATH | 22.4% | — |
| LiveBench Math | 35.5% | — |
| MATH Level 5 | 46.4% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Claude 3.5 Haiku leads
Claude 3.5 Haiku: 18.7 (#281), Gemma 3 4B: 11.8 (#299)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| GPQA Diamond | 38.1% | 23.2% |
| LMArena Expert | 1208 | 1223 |
| MMLU-Pro | 60.5% | — |
| Confabulations | 36.7% | — |
| Vectara Hallucination Rate | — | 6.4% |
| GPQA (HELM) | 36.3% | — |
| MMLU | 74.3% | — |
Multimodal Not comparable
Claude 3.5 Haiku: 26.8 (#117), Gemma 3 4B: —
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Vision | 1092 | — |
| GeoBench | 34% | — |
Multilingual Gemma 3 4B leads
Claude 3.5 Haiku: 40.0 (#218), Gemma 3 4B: 42.5 (#194)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1238 | 1273 |
| LMArena German | 1237 | 1281 |
| LMArena Russian | 1253 | 1294 |
| LMArena Chinese | 1229 | — |
| LMArena French | 1264 | — |
| LMArena Japanese | 1175 | — |
| LMArena Korean | 1173 | — |
| LMArena Spanish | 1261 | — |
Instruction Following Gemma 3 4B leads
Claude 3.5 Haiku: 62.9 (#234), Gemma 3 4B: 65.2 (#225)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1241 | 1239 |
| LiveBench Instruction Following | 61.9% | — |
| IFEval | 79.2% | — |
Long Context Too close to call
Claude 3.5 Haiku: 38.3 (#200), Gemma 3 4B: 38.7 (#194)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1261 | 1273 |
Writing & Preference Too close to call
Claude 3.5 Haiku: 42.7 (#234), Gemma 3 4B: 42.0 (#239)
| Benchmark | Claude 3.5 Haiku | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1255 | 1291 |
| LMArena Creative Writing | 1233 | 1271 |
| EQ-Bench Creative Writing | 1146 | 1068 |
| LMArena Multi-Turn | 1265 | 1255 |
| Short-Story Creative Writing | 73.5% | — |
| WildBench | 76% | — |
| LiveBench Language | 35.4% | — |
Frequently asked questions
Is Claude 3.5 Haiku better than Gemma 3 4B?
Claude 3.5 Haiku is the stronger model overall, scoring 29.2 to 28.1 on the Noometry Index.
Is Claude 3.5 Haiku or Gemma 3 4B better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 32.9 in the Noometry coding category.
How many benchmarks do Claude 3.5 Haiku and Gemma 3 4B share?
17 benchmarks have published results for both models. Claude 3.5 Haiku has 49 scored results on Noometry and Gemma 3 4B has 22.