Model comparison
Llama 3.1-70B vs Qwen1.5-14B
Qwen1.5-14B is the stronger model overall, scoring 32.7 to 29.6 on the Noometry Index.
Last verified . 17 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. Llama 3.1-70B scores higher in 5 categories and Qwen1.5-14B in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen1.5-14B leads 32.4 to 13.5.
Side by side
| Llama 3.1-70B | Qwen1.5-14B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 29.6 | 32.7 |
| Released | 2024-07-23 | 2024-02-04 |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.40 | — |
| Output $ / M tokens | $0.40 | — |
| Results tracked | 35 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen1.5-14B leads
Llama 3.1-70B: 30.3 (#296), Qwen1.5-14B: 33.1 (#263)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Coding | 1260 | 1138 |
| WeirdML | 9% | — |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 54.8% | — |
Agentic & Tool Use Not comparable
Llama 3.1-70B: 25.1 (#112), Qwen1.5-14B: —
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
Reasoning Too close to call
Llama 3.1-70B: 21.6 (#220), Qwen1.5-14B: 21.4 (#223)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1113 |
| DTBench | 60% | — |
| LMCA | 14.8% | — |
| Epoch Capabilities Index | 125.92 | — |
Math Qwen1.5-14B leads
Llama 3.1-70B: 13.5 (#304), Qwen1.5-14B: 32.4 (#215)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Math | 1252 | 1125 |
| OTIS Mock AIME 2024-2025 | 3.6% | — |
| Omni-MATH | 21% | — |
| MATH Level 5 | 36.7% | — |
Knowledge Qwen1.5-14B leads
Llama 3.1-70B: 24.2 (#269), Qwen1.5-14B: 29.8 (#232)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Expert | 1209 | 1094 |
| MMLU | 80.1% | 68.6% |
| GPQA Diamond | 44.2% | — |
| MMLU-Pro | 65.3% | — |
| GPQA (HELM) | 42.6% | — |
Multilingual Llama 3.1-70B leads
Llama 3.1-70B: 38.8 (#225), Qwen1.5-14B: 30.7 (#262)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Non-English | 1219 | 1095 |
| LMArena Chinese | 1215 | 1147 |
| LMArena French | 1261 | 1116 |
| LMArena German | 1222 | 1043 |
| LMArena Japanese | 1132 | 1019 |
| LMArena Russian | 1234 | 1046 |
| LMArena Spanish | 1253 | 1085 |
| LMArena Korean | 1140 | — |
Instruction Following Llama 3.1-70B leads
Llama 3.1-70B: 65.3 (#223), Qwen1.5-14B: 56.8 (#271)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Instruction Following | 1231 | 1102 |
| IFEval | 82.1% | — |
Long Context Llama 3.1-70B leads
Llama 3.1-70B: 37.6 (#214), Qwen1.5-14B: 33.7 (#257)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Longer Query | 1241 | 1113 |
Writing & Preference Llama 3.1-70B leads
Llama 3.1-70B: 35.4 (#267), Qwen1.5-14B: 33.6 (#276)
| Benchmark | Llama 3.1-70B | Qwen1.5-14B |
|---|---|---|
| LMArena Text | 1261 | 1128 |
| LMArena Creative Writing | 1232 | 1091 |
| LMArena Multi-Turn | 1256 | 1110 |
| EQ-Bench Creative Writing | 784 | — |
| WildBench | 75.8% | — |
Frequently asked questions
Is Llama 3.1-70B better than Qwen1.5-14B?
Qwen1.5-14B is the stronger model overall, scoring 32.7 to 29.6 on the Noometry Index.
Is Llama 3.1-70B or Qwen1.5-14B better for coding?
Qwen1.5-14B scores higher on coding benchmarks: 33.1 versus 30.3 in the Noometry coding category.
How many benchmarks do Llama 3.1-70B and Qwen1.5-14B share?
17 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and Qwen1.5-14B has 17.