Model comparison
GPT-4 vs Llama 3-8B
GPT-4 is the stronger model overall, scoring 29.1 to 25.5 on the Noometry Index.
Last verified . 30 shared benchmarks.
MetaSummary
- They share 30 benchmarks with published results for both. GPT-4 scores higher in 7 categories and Llama 3-8B in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-4 leads 18.4 to 7.8.
- The biggest single-benchmark swing is BigCodeBench Complete: 57.2% for GPT-4 and 36.9% for Llama 3-8B.
- Llama 3-8B has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Llama 3-8B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 29.1 | 25.5 |
| Released | 2023-03-14 | 2024-04-18 |
| Weights | Proprietary | Open |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $30 | — |
| Output $ / M tokens | $60 | — |
| Results tracked | 38 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-4: 31.6 (#283), Llama 3-8B: 31.0 (#289)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| BigCodeBench Instruct | 46% | 31.9% |
| LMArena Coding | 1254 | 1152 |
| BigCodeBench Complete | 57.2% | 36.9% |
| HumanEval+ | 79.3% | 56.7% |
| WeirdML | 12.4% | — |
| MBPP+ | — | 54.8% |
Agentic & Tool Use Not comparable
GPT-4: —, Llama 3-8B: —
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning GPT-4 leads
GPT-4: 17.8 (#289), Llama 3-8B: 14.3 (#326)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| Chess Puzzles | 4% | 0% |
| LMArena Hard Prompts | 1241 | 1133 |
| DTBench | 62.7% | 43.9% |
| Epoch Capabilities Index | 125.89 | 116.45 |
| ForecastBench | 57.8 | 58.6 |
| WinoGrande | 87.5% | 75.7% |
| Mystery Game Puzzles | 12% | — |
| LMCA | 17.1% | — |
| Adversarial NLI | — | 57.3% |
| BIG-Bench Hard | 75.1% | — |
| HellaSwag | 95.3% | — |
Math GPT-4 leads
GPT-4: 10.8 (#309), Llama 3-8B: 8.8 (#323)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 1.9% |
| LMArena Math | 1269 | 1151 |
| MATH Level 5 | 23% | 6.1% |
| GSM8K | 92% | — |
Knowledge GPT-4 leads
GPT-4: 18.4 (#282), Llama 3-8B: 7.8 (#308)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 35.7% | 26.1% |
| LMArena Expert | 1211 | 1113 |
| MMLU | 86.4% | 68.8% |
| TriviaQA | 84.8% | 67.7% |
| ARC (AI2) Challenge | — | 82.8% |
| OpenBookQA | — | 82.6% |
Multilingual GPT-4 leads
GPT-4: 40.6 (#215), Llama 3-8B: 30.8 (#261)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1246 | 1098 |
| LMArena Chinese | 1242 | 1076 |
| LMArena French | 1283 | 1159 |
| LMArena German | 1251 | 1104 |
| LMArena Japanese | 1209 | 967 |
| LMArena Korean | 1184 | 1004 |
| LMArena Russian | 1251 | 1109 |
| LMArena Spanish | 1261 | 1173 |
Instruction Following GPT-4 leads
GPT-4: 65.3 (#222), Llama 3-8B: 58.4 (#260)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1241 | 1127 |
Long Context GPT-4 leads
GPT-4: 37.7 (#212), Llama 3-8B: 34.2 (#251)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1244 | 1128 |
Writing & Preference Llama 3-8B leads
GPT-4: 34.9 (#268), Llama 3-8B: 37.5 (#256)
| Benchmark | GPT-4 | Llama 3-8B |
|---|---|---|
| LMArena Text | 1263 | 1166 |
| LMArena Creative Writing | 1244 | 1150 |
| LMArena Multi-Turn | 1257 | 1152 |
| EQ-Bench Creative Writing | 752 | — |
Frequently asked questions
Is GPT-4 better than Llama 3-8B?
GPT-4 is the stronger model overall, scoring 29.1 to 25.5 on the Noometry Index.
Is GPT-4 or Llama 3-8B better for coding?
They score almost the same on coding (31.6 vs 31.0); test both on your own repository before choosing.
How many benchmarks do GPT-4 and Llama 3-8B share?
30 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Llama 3-8B has 34.