Model comparison
GPT-4o mini vs Llama 2-70B
GPT-4o mini is the stronger model overall, scoring 25.4 to 24.1 on the Noometry Index.
Last verified . 26 shared benchmarks.
MetaSummary
- They share 26 benchmarks with published results for both. GPT-4o mini scores higher in 6 categories and Llama 2-70B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where GPT-4o mini leads 42.2 to 27.9.
- The biggest single-benchmark swing is MATH Level 5: 52.6% for GPT-4o mini and 3.3% for Llama 2-70B.
- Llama 2-70B has downloadable open weights; the other is API-only.
Side by side
| GPT-4o mini | Llama 2-70B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 25.4 | 24.1 |
| Released | 2024-07-18 | 2023-07-18 |
| Weights | Proprietary | Open |
| Context window | 128K | — |
| Max output | 16K | — |
| Input $ / M tokens | $0.15 | — |
| Output $ / M tokens | $0.60 | — |
| Results tracked | 60 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-70B leads
GPT-4o mini: 21.6 (#330), Llama 2-70B: 30.8 (#281)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1290 | 1079 |
| 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), Llama 2-70B: —
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| BALROG | 17.4% | — |
Reasoning Llama 2-70B leads
GPT-4o mini: 8.7 (#343), Llama 2-70B: 13.7 (#324)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1267 | 1073 |
| DTBench | 54.4% | 41.6% |
| Epoch Capabilities Index | 126.56 | 113.59 |
| PIQA | 88.7% | 82.8% |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 10.7% | — |
| Kagi LLM Benchmark | 28.8% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 32.8% | — |
| Mystery Game Puzzles | 12% | — |
| LiveBench Data Analysis | 50% | — |
| LMCA | 10.4% | — |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| ForecastBench | — | 51.4 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| LiveBench | 41.3% | — |
| WinoGrande | — | 80.2% |
Math GPT-4o mini leads
GPT-4o mini: 9.9 (#307), Llama 2-70B: 7.5 (#320)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.9% | 0% |
| LMArena Math | 1267 | 1091 |
| MATH Level 5 | 52.6% | 3.3% |
| GSM8K | 91.3% | 69.6% |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| Omni-MATH | 28% | — |
| LiveBench Math | 36.3% | — |
Knowledge GPT-4o mini leads
GPT-4o mini: 17.7 (#279), Llama 2-70B: 7.4 (#305)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| GPQA Diamond | 37.7% | 26.3% |
| LMArena Expert | 1235 | 1039 |
| BoolQ | 88.7% | 88.6% |
| MMLU | 81.8% | 69.9% |
| SimpleQA Verified | 8.3% | — |
| MMLU-Pro | 60.3% | — |
| Confabulations | 37.2% | — |
| GPQA (HELM) | 36.8% | — |
| ARC (AI2) Challenge | — | 78.3% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multimodal Not comparable
GPT-4o mini: 25.3 (#117), Llama 2-70B: —
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
Multilingual GPT-4o mini leads
GPT-4o mini: 42.2 (#194), Llama 2-70B: 27.9 (#269)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1266 | 1045 |
| LMArena Chinese | 1264 | 994 |
| LMArena French | 1296 | 1089 |
| LMArena German | 1272 | 1041 |
| LMArena Japanese | 1216 | 927 |
| LMArena Korean | 1195 | 964 |
| LMArena Russian | 1275 | 1083 |
| LMArena Spanish | 1276 | 1142 |
Instruction Following GPT-4o mini leads
GPT-4o mini: 62.1 (#234), Llama 2-70B: 55.1 (#273)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1259 | 1071 |
| LiveBench Instruction Following | 56.8% | — |
| IFEval | 78.2% | — |
Long Context GPT-4o mini leads
GPT-4o mini: 39.1 (#181), Llama 2-70B: 32.2 (#265)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1289 | 1062 |
Writing & Preference GPT-4o mini leads
GPT-4o mini: 39.7 (#243), Llama 2-70B: 32.5 (#274)
| Benchmark | GPT-4o mini | Llama 2-70B |
|---|---|---|
| LMArena Text | 1286 | 1115 |
| LMArena Creative Writing | 1268 | 1075 |
| LMArena Multi-Turn | 1285 | 1088 |
| 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 Llama 2-70B?
GPT-4o mini is the stronger model overall, scoring 25.4 to 24.1 on the Noometry Index.
Is GPT-4o mini or Llama 2-70B better for coding?
Llama 2-70B scores higher on coding benchmarks: 30.8 versus 21.6 in the Noometry coding category.
How many benchmarks do GPT-4o mini and Llama 2-70B share?
26 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Llama 2-70B has 35.