Model comparison
GPT-4 vs Qwen-14B
Qwen-14B is the stronger model overall, scoring 31.4 to 29.1 on the Noometry Index.
Last verified . 14 shared benchmarks.
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. GPT-4 scores higher in 5 categories and Qwen-14B in 2 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen-14B leads 31.2 to 10.8.
- Qwen-14B has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Qwen-14B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 29.1 | 31.4 |
| Released | 2023-03-14 | 2023-09-24 |
| Weights | Proprietary | Open |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $30 | — |
| Output $ / M tokens | $60 | — |
| Results tracked | 38 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-4: 31.6 (#283), Qwen-14B: 31.2 (#288)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Coding | 1254 | 1071 |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Qwen-14B: —
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning Qwen-14B leads
GPT-4: 17.8 (#289), Qwen-14B: 19.6 (#257)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1027 |
| BIG-Bench Hard | 75.1% | 55% |
| Epoch Capabilities Index | 125.89 | 113.03 |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| LAMBADA | — | 71.1% |
| PIQA | — | 79.9% |
| WinoGrande | 87.5% | — |
Math Qwen-14B leads
GPT-4: 10.8 (#309), Qwen-14B: 31.2 (#227)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Math | 1269 | 1068 |
| GSM8K | 92% | 61.3% |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| MATH Level 5 | 23% | — |
Knowledge Not comparable
GPT-4: 18.4 (#282), Qwen-14B: —
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| MMLU | 86.4% | 66.3% |
| GPQA Diamond | 35.7% | — |
| LMArena Expert | 1211 | — |
| ARC (AI2) Challenge | — | 84.4% |
| BoolQ | — | 86.2% |
| TriviaQA | 84.8% | — |
Multilingual GPT-4 leads
GPT-4: 40.6 (#215), Qwen-14B: 27.5 (#275)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Non-English | 1246 | 1041 |
| LMArena Chinese | 1242 | 1077 |
| LMArena French | 1283 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1209 | — |
| LMArena Korean | 1184 | — |
| LMArena Russian | 1251 | — |
| LMArena Spanish | 1261 | — |
Instruction Following GPT-4 leads
GPT-4: 65.3 (#222), Qwen-14B: 52.4 (#289)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Instruction Following | 1241 | 1031 |
Long Context GPT-4 leads
GPT-4: 37.7 (#212), Qwen-14B: 31.3 (#280)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Longer Query | 1244 | 1028 |
Writing & Preference GPT-4 leads
GPT-4: 34.9 (#268), Qwen-14B: 27.6 (#299)
| Benchmark | GPT-4 | Qwen-14B |
|---|---|---|
| LMArena Text | 1263 | 1051 |
| LMArena Creative Writing | 1244 | 1028 |
| LMArena Multi-Turn | 1257 | 1022 |
| EQ-Bench Creative Writing | 752 | — |
Frequently asked questions
Is GPT-4 better than Qwen-14B?
Qwen-14B is the stronger model overall, scoring 31.4 to 29.1 on the Noometry Index.
Is GPT-4 or Qwen-14B better for coding?
They score almost the same on coding (31.6 vs 31.2); test both on your own repository before choosing.
How many benchmarks do GPT-4 and Qwen-14B share?
14 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Qwen-14B has 18.