Model comparison
Llama 3.1-8B vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 22.8 on the Noometry Index. Llama 3.1-8B costs 43× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 34 benchmarks with published results for both. Llama 3.1-8B scores higher in 1 category and Qwen2.5 72B Instruct in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen2.5 72B Instruct leads 26.9 to 8.0.
- The biggest single-benchmark swing is MATH Level 5: 22.9% for Llama 3.1-8B and 63.2% for Qwen2.5 72B Instruct.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct accepts more context: 131K tokens versus 128K.
Side by side
| Llama 3.1-8B | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 22.8 | 31.5 |
| Released | 2024-07-23 | 2024-09 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 4K | 8K |
| Input $ / M tokens | $0.05 | $1.40 |
| Output $ / M tokens | $0.08 | $5.60 |
| Results tracked | 43 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5 72B Instruct leads
Llama 3.1-8B: 19.9 (#335), Qwen2.5 72B Instruct: 32.8 (#256)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| WeirdML | 1.7% | 16% |
| BigCodeBench Instruct | 32.8% | 45.8% |
| LMArena Coding | 1195 | 1293 |
| BigCodeBench Complete | 40.5% | 55.9% |
| SciCode | 13.2% | — |
| HumanEval+ | 62.8% | — |
| MBPP+ | 55.6% | — |
Agentic & Tool Use Too close to call
Llama 3.1-8B: 22.3 (#123), Qwen2.5 72B Instruct: 22.0 (#124)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| BALROG | 15.1% | 16.2% |
| Berkeley Function Calling Leaderboard | 25.8% | — |
| TheAgentCompany | — | 5.7% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Llama 3.1-8B: 14.5 (#316), Qwen2.5 72B Instruct: 21.4 (#197)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1175 | 1271 |
| DTBench | 50.9% | 62.9% |
| LMCA | 5.4% | 13.4% |
| Epoch Capabilities Index | 116.48 | 129.01 |
| PIQA | 81.2% | 82.6% |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| BIG-Bench Hard | — | 79.8% |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| WinoGrande | — | 82.3% |
Math Qwen2.5 72B Instruct leads
Llama 3.1-8B: 9.5 (#311), Qwen2.5 72B Instruct: 18.1 (#281)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.7% | 8.1% |
| Omni-MATH | 13.7% | 33% |
| LMArena Math | 1179 | 1283 |
| MATH Level 5 | 22.9% | 63.2% |
| GSM8K | 82.4% | — |
Knowledge Qwen2.5 72B Instruct leads
Llama 3.1-8B: 8.0 (#302), Qwen2.5 72B Instruct: 26.9 (#247)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 27% | 49.1% |
| MMLU-Pro | 40.6% | 63.1% |
| GPQA (HELM) | 24.7% | 42.6% |
| LMArena Expert | 1144 | 1245 |
| MMLU | 56.1% | 85.3% |
| Confabulations | — | 19.1% |
| ARC (AI2) Challenge | — | 94.5% |
| BoolQ | 82.8% | — |
| TriviaQA | — | 71.9% |
Multilingual Qwen2.5 72B Instruct leads
Llama 3.1-8B: 34.2 (#244), Qwen2.5 72B Instruct: 41.2 (#208)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1148 | 1252 |
| LMArena Chinese | 1151 | 1272 |
| LMArena French | 1176 | 1279 |
| LMArena German | 1144 | 1234 |
| LMArena Japanese | 1061 | 1180 |
| LMArena Korean | 1053 | 1188 |
| LMArena Russian | 1158 | 1264 |
| LMArena Spanish | 1169 | 1255 |
Instruction Following Qwen2.5 72B Instruct leads
Llama 3.1-8B: 59.1 (#253), Qwen2.5 72B Instruct: 65.7 (#216)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| IFEval | 74.3% | 80.6% |
| LMArena Instruction Following | 1159 | 1254 |
Long Context Qwen2.5 72B Instruct leads
Llama 3.1-8B: 35.8 (#233), Qwen2.5 72B Instruct: 38.9 (#183)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1182 | 1282 |
Writing & Preference Qwen2.5 72B Instruct leads
Llama 3.1-8B: 29.9 (#285), Qwen2.5 72B Instruct: 46.9 (#210)
| Benchmark | Llama 3.1-8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1187 | 1269 |
| LMArena Creative Writing | 1154 | 1221 |
| WildBench | 68.7% | 80.2% |
| LMArena Multi-Turn | 1172 | 1272 |
| EQ-Bench Creative Writing | 713 | — |
Frequently asked questions
Is Llama 3.1-8B better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 22.8 on the Noometry Index. Llama 3.1-8B costs 43× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Which is cheaper, Llama 3.1-8B or Qwen2.5 72B Instruct?
Llama 3.1-8B is cheaper. It lists at $0.05 per million input tokens and $0.08 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Llama 3.1-8B or Qwen2.5 72B Instruct better for coding?
Qwen2.5 72B Instruct scores higher on coding benchmarks: 32.8 versus 19.9 in the Noometry coding category.
Which has the bigger context window?
Qwen2.5 72B Instruct does, with 131K tokens against 128K.
How many benchmarks do Llama 3.1-8B and Qwen2.5 72B Instruct share?
34 benchmarks have published results for both models. Llama 3.1-8B has 43 scored results on Noometry and Qwen2.5 72B Instruct has 43.