Model comparison
Codellama 70b Instruct vs Qwen-14B
Codellama 70b Instruct is the stronger model overall, scoring 33.3 to 30.8 on the Noometry Index.
Last verified . 4 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 4 benchmarks with published results for both. Codellama 70b Instruct scores higher in 3 categories and Qwen-14B in 2 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in coding, where Codellama 70b Instruct leads 37.0 to 30.6.
Side by side
| Codellama 70b Instruct | Qwen-14B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 33.3 | 30.8 |
| Released | — | 2023-09-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 7 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Codellama 70b Instruct leads
Codellama 70b Instruct: 37.0 (#189), Qwen-14B: 30.6 (#283)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| BigCodeBench Instruct | 40.7% | — |
| LMArena Coding | — | 1072 |
| BigCodeBench Complete | 49.6% | — |
| HumanEval+ | 65.9% | — |
Reasoning Too close to call
Codellama 70b Instruct: 19.2 (#241), Qwen-14B: 18.7 (#255)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Hard Prompts | 1052 | 1027 |
| BIG-Bench Hard | — | 55% |
| Epoch Capabilities Index | — | 112.8 |
| LAMBADA | — | 71.1% |
| PIQA | — | 79.9% |
Math Not comparable
Codellama 70b Instruct: —, Qwen-14B: 28.7 (#224)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Math | — | 1068 |
| GSM8K | — | 61.3% |
Knowledge Not comparable
Codellama 70b Instruct: —, Qwen-14B: —
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| ARC (AI2) Challenge | — | 84.4% |
| BoolQ | — | 86.2% |
| MMLU | — | 66.3% |
Multilingual Qwen-14B leads
Codellama 70b Instruct: 24.9 (#283), Qwen-14B: 27.7 (#270)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Non-English | 992 | 1042 |
| LMArena Chinese | — | 1077 |
Instruction Following Too close to call
Codellama 70b Instruct: 52.1 (#288), Qwen-14B: 52.6 (#284)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Instruction Following | 1024 | 1031 |
Long Context Not comparable
Codellama 70b Instruct: —, Qwen-14B: 31.3 (#276)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Longer Query | — | 1029 |
Writing & Preference Codellama 70b Instruct leads
Codellama 70b Instruct: 33.6 (#272), Qwen-14B: 27.8 (#294)
| Benchmark | Codellama 70b Instruct | Qwen-14B |
|---|---|---|
| LMArena Text | 1057 | 1051 |
| LMArena Creative Writing | — | 1028 |
| LMArena Multi-Turn | — | 1022 |
Frequently asked questions
Is Codellama 70b Instruct better than Qwen-14B?
Codellama 70b Instruct is the stronger model overall, scoring 33.3 to 30.8 on the Noometry Index.
Is Codellama 70b Instruct or Qwen-14B better for coding?
Codellama 70b Instruct scores higher on coding benchmarks: 37.0 versus 30.6 in the Noometry coding category.
How many benchmarks do Codellama 70b Instruct and Qwen-14B share?
4 benchmarks have published results for both models. Codellama 70b Instruct has 7 scored results on Noometry and Qwen-14B has 18.