Model comparison
GPT-4 vs Llama 13b
GPT-4 is the stronger model overall, scoring 29.1 to 24.4 on the Noometry Index.
Last verified . 15 shared benchmarks.
MetaSummary
- They share 15 benchmarks with published results for both. GPT-4 scores higher in 5 categories and Llama 13b in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where GPT-4 leads 65.3 to 36.7.
- Llama 13b has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Llama 13b | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 29.1 | 24.4 |
| Released | 2023-03-14 | 2023-02-24 |
| Weights | Proprietary | Open |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $30 | — |
| Output $ / M tokens | $60 | — |
| Results tracked | 38 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4 leads
GPT-4: 31.6 (#283), Llama 13b: 21.4 (#337)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Coding | 1254 | 683 |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Llama 13b: —
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning GPT-4 leads
GPT-4: 17.8 (#289), Llama 13b: 14.0 (#329)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Hard Prompts | 1241 | 728 |
| BIG-Bench Hard | 75.1% | 37.9% |
| Epoch Capabilities Index | 125.89 | 100.58 |
| HellaSwag | 95.3% | 79.2% |
| WinoGrande | 87.5% | 73% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| ForecastBench | 57.8 | — |
| LAMBADA | — | 75.2% |
| PIQA | — | 80.1% |
Math Llama 13b leads
GPT-4: 10.8 (#309), Llama 13b: 26.7 (#256)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Math | 1269 | 838 |
| GSM8K | 92% | 20.6% |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| MATH Level 5 | 23% | — |
Knowledge Not comparable
GPT-4: 18.4 (#282), Llama 13b: —
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| MMLU | 86.4% | 47.7% |
| TriviaQA | 84.8% | 77.9% |
| GPQA Diamond | 35.7% | — |
| LMArena Expert | 1211 | — |
| ARC (AI2) Challenge | — | 52.7% |
| BoolQ | — | 78.7% |
| OpenBookQA | — | 56.4% |
Multimodal Not comparable
GPT-4: —, Llama 13b: —
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual GPT-4 leads
GPT-4: 40.6 (#215), Llama 13b: 16.6 (#297)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Non-English | 1246 | 819 |
| LMArena Chinese | 1242 | — |
| LMArena French | 1283 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1209 | — |
| LMArena Korean | 1184 | — |
| LMArena Russian | 1251 | — |
| LMArena Spanish | 1261 | — |
Instruction Following GPT-4 leads
GPT-4: 65.3 (#222), Llama 13b: 36.7 (#305)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Instruction Following | 1241 | 781 |
Long Context Not comparable
GPT-4: 37.7 (#212), Llama 13b: —
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Longer Query | 1244 | — |
Writing & Preference GPT-4 leads
GPT-4: 34.9 (#268), Llama 13b: 13.8 (#312)
| Benchmark | GPT-4 | Llama 13b |
|---|---|---|
| LMArena Text | 1263 | 834 |
| LMArena Creative Writing | 1244 | 794 |
| LMArena Multi-Turn | 1257 | 753 |
| EQ-Bench Creative Writing | 752 | — |
Frequently asked questions
Is GPT-4 better than Llama 13b?
GPT-4 is the stronger model overall, scoring 29.1 to 24.4 on the Noometry Index.
Is GPT-4 or Llama 13b better for coding?
GPT-4 scores higher on coding benchmarks: 31.6 versus 21.4 in the Noometry coding category.
How many benchmarks do GPT-4 and Llama 13b share?
15 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Llama 13b has 21.