Model comparison
Gemma 3 4B vs Tulu 3 (Tülu 3) 70B
Tulu 3 (Tülu 3) 70B is the stronger model overall, scoring 33.0 to 28.1 on the Noometry Index.
Last verified . 12 shared benchmarks.
Google
Allen Institute for AI (Ai2)Summary
- They share 12 benchmarks with published results for both. Gemma 3 4B scores higher in 4 categories and Tulu 3 (Tülu 3) 70B in 4 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Tulu 3 (Tülu 3) 70B leads 25.0 to 11.8.
- The biggest single-benchmark swing is GPQA Diamond: 23.2% for Gemma 3 4B and 46.3% for Tulu 3 (Tülu 3) 70B.
Side by side
| Gemma 3 4B | Tulu 3 (Tülu 3) 70B | |
|---|---|---|
| Provider | Allen Institute for AI (Ai2) | |
| Noometry Index | 28.1 | 33.0 |
| Released | 2025-03-12 | 2024-11-21 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.04 | — |
| Output $ / M tokens | $0.08 | — |
| Results tracked | 22 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 3 4B: 35.9 (#215), Tulu 3 (Tülu 3) 70B: 36.0 (#214)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Coding | 1230 | 1235 |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.9 (#142), Tulu 3 (Tülu 3) 70B: —
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
Reasoning Tulu 3 (Tülu 3) 70B leads
Gemma 3 4B: 13.2 (#335), Tulu 3 (Tülu 3) 70B: 23.9 (#169)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Hard Prompts | 1253 | 1220 |
| Kagi LLM Benchmark | 25.2% | — |
| Chess Puzzles | 0% | — |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| Epoch Capabilities Index | 116.02 | — |
Math Gemma 3 4B leads
Gemma 3 4B: 16.8 (#292), Tulu 3 (Tülu 3) 70B: 14.2 (#303)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 4.4% |
| LMArena Math | 1239 | 1242 |
| MATH Level 5 | — | 42.7% |
Knowledge Tulu 3 (Tülu 3) 70B leads
Gemma 3 4B: 11.8 (#299), Tulu 3 (Tülu 3) 70B: 25.0 (#264)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| GPQA Diamond | 23.2% | 46.3% |
| Vectara Hallucination Rate | 6.4% | — |
| LMArena Expert | 1223 | — |
Multilingual Gemma 3 4B leads
Gemma 3 4B: 42.5 (#194), Tulu 3 (Tülu 3) 70B: 39.9 (#222)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Non-English | 1273 | 1236 |
| LMArena Russian | 1294 | 1246 |
| LMArena Chinese | — | 1249 |
| LMArena German | 1281 | — |
Instruction Following Too close to call
Gemma 3 4B: 65.2 (#225), Tulu 3 (Tülu 3) 70B: 64.8 (#227)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Instruction Following | 1239 | 1233 |
Long Context Gemma 3 4B leads
Gemma 3 4B: 38.7 (#194), Tulu 3 (Tülu 3) 70B: 37.1 (#222)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Longer Query | 1273 | 1224 |
Writing & Preference Tulu 3 (Tülu 3) 70B leads
Gemma 3 4B: 42.0 (#239), Tulu 3 (Tülu 3) 70B: 45.6 (#223)
| Benchmark | Gemma 3 4B | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Text | 1291 | 1256 |
| LMArena Creative Writing | 1271 | 1231 |
| LMArena Multi-Turn | 1255 | 1252 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Tulu 3 (Tülu 3) 70B?
Tulu 3 (Tülu 3) 70B is the stronger model overall, scoring 33.0 to 28.1 on the Noometry Index.
Is Gemma 3 4B or Tulu 3 (Tülu 3) 70B better for coding?
They score almost the same on coding (35.9 vs 36.0); test both on your own repository before choosing.
How many benchmarks do Gemma 3 4B and Tulu 3 (Tülu 3) 70B share?
12 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Tulu 3 (Tülu 3) 70B has 14.