Model comparison
GPT-4o mini vs Tulu 3 (Tülu 3) 70B
Tulu 3 (Tülu 3) 70B is the stronger model overall, scoring 32.6 to 25.4 on the Noometry Index.
Last verified . 14 shared benchmarks.
Allen Institute for AI (Ai2)Summary
- They share 14 benchmarks with published results for both. GPT-4o mini scores higher in 2 categories and Tulu 3 (Tülu 3) 70B in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Tulu 3 (Tülu 3) 70B leads 22.9 to 8.7.
- The biggest single-benchmark swing is MATH Level 5: 52.6% for GPT-4o mini and 42.7% for Tulu 3 (Tülu 3) 70B.
- Tulu 3 (Tülu 3) 70B has downloadable open weights; the other is API-only.
Side by side
| GPT-4o mini | Tulu 3 (Tülu 3) 70B | |
|---|---|---|
| Provider | OpenAI | Allen Institute for AI (Ai2) |
| Noometry Index | 25.4 | 32.6 |
| Released | 2024-07-18 | 2024-11-21 |
| Weights | Proprietary | Open |
| Context window | 128K | — |
| Max output | 16K | — |
| Input $ / M tokens | $0.15 | — |
| Output $ / M tokens | $0.60 | — |
| Results tracked | 60 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 21.6 (#330), Tulu 3 (Tülu 3) 70B: 35.4 (#211)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Coding | 1290 | 1236 |
| Aider Polyglot | 3.6% | — |
| WeirdML | 11.8% | — |
| BigCodeBench Instruct | 46.1% | — |
| LiveBench Coding | 43.1% | — |
| BigCodeBench Complete | 57.4% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use Not comparable
GPT-4o mini: 27.2 (#95), Tulu 3 (Tülu 3) 70B: —
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| BALROG | 17.4% | — |
Reasoning Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 8.7 (#343), Tulu 3 (Tülu 3) 70B: 22.9 (#172)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Hard Prompts | 1267 | 1220 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 10.7% | — |
| Kagi LLM Benchmark | 28.8% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 32.8% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 54.4% | — |
| LiveBench Data Analysis | 50% | — |
| LMCA | 10.4% | — |
| Epoch Capabilities Index | 126.56 | — |
| LiveBench | 41.3% | — |
| PIQA | 88.7% | — |
Math Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 9.9 (#307), Tulu 3 (Tülu 3) 70B: 13.3 (#297)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.9% | 4.4% |
| LMArena Math | 1267 | 1242 |
| MATH Level 5 | 52.6% | 42.7% |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| Omni-MATH | 28% | — |
| LiveBench Math | 36.3% | — |
| GSM8K | 91.3% | — |
Knowledge Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 17.7 (#279), Tulu 3 (Tülu 3) 70B: 25.1 (#257)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| GPQA Diamond | 37.7% | 46.3% |
| SimpleQA Verified | 8.3% | — |
| MMLU-Pro | 60.3% | — |
| Confabulations | 37.2% | — |
| GPQA (HELM) | 36.8% | — |
| LMArena Expert | 1235 | — |
| BoolQ | 88.7% | — |
| MMLU | 81.8% | — |
Multimodal Not comparable
GPT-4o mini: 25.3 (#117), Tulu 3 (Tülu 3) 70B: —
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
Multilingual GPT-4o mini leads
GPT-4o mini: 42.2 (#194), Tulu 3 (Tülu 3) 70B: 40.1 (#217)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Non-English | 1266 | 1236 |
| LMArena Chinese | 1264 | 1249 |
| LMArena Russian | 1275 | 1246 |
| LMArena French | 1296 | — |
| LMArena German | 1272 | — |
| LMArena Japanese | 1216 | — |
| LMArena Korean | 1195 | — |
| LMArena Spanish | 1276 | — |
Instruction Following Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 62.1 (#234), Tulu 3 (Tülu 3) 70B: 65.0 (#222)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Instruction Following | 1259 | 1233 |
| LiveBench Instruction Following | 56.8% | — |
| IFEval | 78.2% | — |
Long Context GPT-4o mini leads
GPT-4o mini: 39.1 (#181), Tulu 3 (Tülu 3) 70B: 37.1 (#217)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Longer Query | 1289 | 1224 |
Writing & Preference Tulu 3 (Tülu 3) 70B leads
GPT-4o mini: 39.7 (#243), Tulu 3 (Tülu 3) 70B: 45.8 (#218)
| Benchmark | GPT-4o mini | Tulu 3 (Tülu 3) 70B |
|---|---|---|
| LMArena Text | 1286 | 1256 |
| LMArena Creative Writing | 1268 | 1231 |
| LMArena Multi-Turn | 1285 | 1252 |
| Short-Story Creative Writing | 67.2% | — |
| EQ-Bench Creative Writing | 873 | — |
| WildBench | 79.1% | — |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than Tulu 3 (Tülu 3) 70B?
Tulu 3 (Tülu 3) 70B is the stronger model overall, scoring 32.6 to 25.4 on the Noometry Index.
Is GPT-4o mini or Tulu 3 (Tülu 3) 70B better for coding?
Tulu 3 (Tülu 3) 70B scores higher on coding benchmarks: 35.4 versus 21.6 in the Noometry coding category.
How many benchmarks do GPT-4o mini and Tulu 3 (Tülu 3) 70B share?
14 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Tulu 3 (Tülu 3) 70B has 14.