Model comparison
GPT-4 vs Llama 3.2 1B
GPT-4 is the stronger model overall, scoring 29.1 to 20.1 on the Noometry Index. Llama 3.2 1B costs 532× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
MetaSummary
- They share 20 benchmarks with published results for both. GPT-4 scores higher in 8 categories and Llama 3.2 1B in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where GPT-4 leads 40.6 to 23.8.
- The biggest single-benchmark swing is BigCodeBench Complete: 57.2% for GPT-4 and 11.3% for Llama 3.2 1B.
- Llama 3.2 1B is cheaper at $0.027 / $0.20 per million input/output tokens, against $30 / $60 for GPT-4.
- Llama 3.2 1B accepts more context: 60K tokens versus 8K.
- Llama 3.2 1B has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Llama 3.2 1B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 29.1 | 20.1 |
| Released | 2023-03-14 | 2024-09-24 |
| Weights | Proprietary | Open |
| Context window | 8K | 60K |
| Max output | 8K | 54K |
| Input $ / M tokens | $30 | $0.027 |
| Output $ / M tokens | $60 | $0.20 |
| Results tracked | 38 | 22 |
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 3.2 1B: 21.1 (#338)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| BigCodeBench Instruct | 46% | 8.2% |
| LMArena Coding | 1254 | 1070 |
| BigCodeBench Complete | 57.2% | 11.3% |
| WeirdML | 12.4% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Llama 3.2 1B: 14.6 (#150)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 10.8% |
| BALROG | — | 6.6% |
| METR Time Horizons | 36.1% | — |
Reasoning GPT-4 leads
GPT-4: 17.8 (#289), Llama 3.2 1B: 16.2 (#308)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| Chess Puzzles | 4% | 0% |
| LMArena Hard Prompts | 1241 | 1044 |
| Epoch Capabilities Index | 125.89 | 101.99 |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math Too close to call
GPT-4: 10.8 (#309), Llama 3.2 1B: 10.4 (#313)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 0.6% |
| LMArena Math | 1269 | 1086 |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |
Knowledge GPT-4 leads
GPT-4: 18.4 (#282), Llama 3.2 1B: 7.2 (#312)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| GPQA Diamond | 35.7% | 23.9% |
| LMArena Expert | 1211 | 1007 |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multilingual GPT-4 leads
GPT-4: 40.6 (#215), Llama 3.2 1B: 23.8 (#292)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| LMArena Non-English | 1246 | 973 |
| LMArena Chinese | 1242 | 959 |
| LMArena German | 1251 | 1014 |
| LMArena Russian | 1251 | 941 |
| LMArena French | 1283 | — |
| LMArena Japanese | 1209 | — |
| LMArena Korean | 1184 | — |
| LMArena Spanish | 1261 | — |
Instruction Following GPT-4 leads
GPT-4: 65.3 (#222), Llama 3.2 1B: 52.4 (#290)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| LMArena Instruction Following | 1241 | 1031 |
Long Context GPT-4 leads
GPT-4: 37.7 (#212), Llama 3.2 1B: 31.9 (#274)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| LMArena Longer Query | 1244 | 1050 |
Writing & Preference GPT-4 leads
GPT-4: 34.9 (#268), Llama 3.2 1B: 21.3 (#310)
| Benchmark | GPT-4 | Llama 3.2 1B |
|---|---|---|
| LMArena Text | 1263 | 1055 |
| LMArena Creative Writing | 1244 | 1033 |
| EQ-Bench Creative Writing | 752 | 200 |
| LMArena Multi-Turn | 1257 | 1030 |
Frequently asked questions
Is GPT-4 better than Llama 3.2 1B?
GPT-4 is the stronger model overall, scoring 29.1 to 20.1 on the Noometry Index. Llama 3.2 1B costs 532× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.
Which is cheaper, GPT-4 or Llama 3.2 1B?
Llama 3.2 1B is cheaper. It lists at $0.027 per million input tokens and $0.20 per million output tokens; GPT-4 lists at $30 and $60.
Is GPT-4 or Llama 3.2 1B better for coding?
GPT-4 scores higher on coding benchmarks: 31.6 versus 21.1 in the Noometry coding category.
Which has the bigger context window?
Llama 3.2 1B does, with 60K tokens against 8K.
How many benchmarks do GPT-4 and Llama 3.2 1B share?
20 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Llama 3.2 1B has 22.