Model comparison
GPT-4 vs Yi-Lightning
Yi-Lightning is the stronger model overall, scoring 37.1 to 29.1 on the Noometry Index.
Last verified . 17 shared benchmarks.
Summary
- They share 17 benchmarks with published results for both. GPT-4 scores higher in 1 category and Yi-Lightning in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Yi-Lightning leads 36.2 to 10.8.
Side by side
| GPT-4 | Yi-Lightning | |
|---|---|---|
| Provider | OpenAI | 01.AI |
| Noometry Index | 29.1 | 37.1 |
| Released | 2023-03-14 | 2024-12-02 |
| Weights | Proprietary | Proprietary |
| 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 GPT-4 leads
GPT-4: 31.6 (#283), Yi-Lightning: 27.4 (#320)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Coding | 1254 | 1312 |
| Aider Polyglot | — | 12.9% |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Yi-Lightning: —
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning Yi-Lightning leads
GPT-4: 17.8 (#289), Yi-Lightning: 25.9 (#139)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1302 |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| Epoch Capabilities Index | 125.89 | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math Yi-Lightning leads
GPT-4: 10.8 (#309), Yi-Lightning: 36.2 (#172)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Math | 1269 | 1300 |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |
Knowledge Yi-Lightning leads
GPT-4: 18.4 (#282), Yi-Lightning: 35.4 (#185)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Expert | 1211 | 1286 |
| GPQA Diamond | 35.7% | — |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multilingual Yi-Lightning leads
GPT-4: 40.6 (#215), Yi-Lightning: 42.2 (#196)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Non-English | 1246 | 1269 |
| LMArena Chinese | 1242 | 1320 |
| LMArena French | 1283 | 1305 |
| LMArena German | 1251 | 1267 |
| LMArena Japanese | 1209 | 1228 |
| LMArena Korean | 1184 | 1192 |
| LMArena Russian | 1251 | 1255 |
| LMArena Spanish | 1261 | 1315 |
Instruction Following Yi-Lightning leads
GPT-4: 65.3 (#222), Yi-Lightning: 67.4 (#196)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Instruction Following | 1241 | 1278 |
Long Context Yi-Lightning leads
GPT-4: 37.7 (#212), Yi-Lightning: 39.4 (#181)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Longer Query | 1244 | 1297 |
Writing & Preference Yi-Lightning leads
GPT-4: 34.9 (#268), Yi-Lightning: 50.2 (#184)
| Benchmark | GPT-4 | Yi-Lightning |
|---|---|---|
| LMArena Text | 1263 | 1302 |
| LMArena Creative Writing | 1244 | 1280 |
| LMArena Multi-Turn | 1257 | 1311 |
| EQ-Bench Creative Writing | 752 | — |
Frequently asked questions
Is GPT-4 better than Yi-Lightning?
Yi-Lightning is the stronger model overall, scoring 37.1 to 29.1 on the Noometry Index.
Is GPT-4 or Yi-Lightning better for coding?
GPT-4 scores higher on coding benchmarks: 31.6 versus 27.4 in the Noometry coding category.
How many benchmarks do GPT-4 and Yi-Lightning share?
17 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Yi-Lightning has 18.