Model comparison
Llama 3-70B vs Llama 3-8B
Llama 3-70B is the stronger model overall, scoring 28.8 to 25.5 on the Noometry Index.
Last verified . 29 shared benchmarks.
MetaSummary
- They share 29 benchmarks with published results for both. Llama 3-70B scores higher in 8 categories and Llama 3-8B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3-70B leads 20.8 to 7.8.
- The biggest single-benchmark swing is BigCodeBench Complete: 54.5% for Llama 3-70B and 36.9% for Llama 3-8B.
Side by side
| Llama 3-70B | Llama 3-8B | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 28.8 | 25.5 |
| Released | 2024-04-18 | 2024-04-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 31 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-70B leads
Llama 3-70B: 35.8 (#218), Llama 3-8B: 31.0 (#289)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| BigCodeBench Instruct | 43.6% | 31.9% |
| LMArena Coding | 1206 | 1152 |
| BigCodeBench Complete | 54.5% | 36.9% |
| HumanEval+ | 72% | 56.7% |
| MBPP+ | 69% | 54.8% |
Agentic & Tool Use Not comparable
Llama 3-70B: 21.1 (#139), Llama 3-8B: —
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| Cybench | 5% | — |
Reasoning Llama 3-70B leads
Llama 3-70B: 18.0 (#288), Llama 3-8B: 14.3 (#326)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| LMArena Hard Prompts | 1195 | 1133 |
| DTBench | 54.2% | 43.9% |
| Epoch Capabilities Index | 122.93 | 116.45 |
| ForecastBench | 57.1 | 58.6 |
| WinoGrande | 83.5% | 75.7% |
| Kagi LLM Benchmark | 35.1% | — |
| Chess Puzzles | — | 0% |
| Adversarial NLI | — | 57.3% |
Math Llama 3-70B leads
Llama 3-70B: 12.8 (#305), Llama 3-8B: 8.8 (#323)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 1.9% |
| LMArena Math | 1218 | 1151 |
| MATH Level 5 | 22.6% | 6.1% |
Knowledge Llama 3-70B leads
Llama 3-70B: 20.8 (#277), Llama 3-8B: 7.8 (#308)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 40.6% | 26.1% |
| LMArena Expert | 1149 | 1113 |
| MMLU | 79.3% | 68.8% |
| ARC (AI2) Challenge | — | 82.8% |
| OpenBookQA | — | 82.6% |
| TriviaQA | — | 67.7% |
Multilingual Llama 3-70B leads
Llama 3-70B: 33.6 (#251), Llama 3-8B: 30.8 (#261)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1142 | 1098 |
| LMArena Chinese | 1114 | 1076 |
| LMArena French | 1232 | 1159 |
| LMArena German | 1169 | 1104 |
| LMArena Japanese | 1017 | 967 |
| LMArena Korean | 1017 | 1004 |
| LMArena Russian | 1159 | 1109 |
| LMArena Spanish | 1241 | 1173 |
Instruction Following Llama 3-70B leads
Llama 3-70B: 62.5 (#238), Llama 3-8B: 58.4 (#260)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1194 | 1127 |
Long Context Llama 3-70B leads
Llama 3-70B: 35.6 (#240), Llama 3-8B: 34.2 (#251)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1174 | 1128 |
Writing & Preference Llama 3-70B leads
Llama 3-70B: 42.8 (#231), Llama 3-8B: 37.5 (#256)
| Benchmark | Llama 3-70B | Llama 3-8B |
|---|---|---|
| LMArena Text | 1221 | 1166 |
| LMArena Creative Writing | 1210 | 1150 |
| LMArena Multi-Turn | 1223 | 1152 |
Frequently asked questions
Is Llama 3-70B better than Llama 3-8B?
Llama 3-70B is the stronger model overall, scoring 28.8 to 25.5 on the Noometry Index.
Is Llama 3-70B or Llama 3-8B better for coding?
Llama 3-70B scores higher on coding benchmarks: 35.8 versus 31.0 in the Noometry coding category.
How many benchmarks do Llama 3-70B and Llama 3-8B share?
29 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and Llama 3-8B has 34.