Model comparison
GPT-4 Turbo vs Phi 3 Small 8k Instruct
GPT-4 Turbo is the stronger model overall, scoring 30.5 to 29.3 on the Noometry Index.
Last verified . 18 shared benchmarks.
MicrosoftSummary
- They share 18 benchmarks with published results for both. GPT-4 Turbo scores higher in 6 categories and Phi 3 Small 8k Instruct in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Phi 3 Small 8k Instruct leads 27.6 to 9.0.
- Phi 3 Small 8k Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-4 Turbo | Phi 3 Small 8k Instruct | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 30.5 | 29.3 |
| Released | 2023-11-06 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | 128K | — |
| Max output | 4K | — |
| Input $ / M tokens | $10 | — |
| Output $ / M tokens | $30 | — |
| Results tracked | 36 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4 Turbo leads
GPT-4 Turbo: 33.8 (#249), Phi 3 Small 8k Instruct: 27.9 (#318)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Coding | 1268 | 1101 |
| WeirdML | 18% | — |
| BigCodeBench Instruct | 48.2% | — |
| LiveBench Coding | — | 20.3% |
| BigCodeBench Complete | 58.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73.3% | — |
Agentic & Tool Use Not comparable
GPT-4 Turbo: —, Phi 3 Small 8k Instruct: —
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| METR Time Horizons | 36.7% | — |
Reasoning Too close to call
GPT-4 Turbo: 15.3 (#317), Phi 3 Small 8k Instruct: 14.8 (#323)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1100 |
| SimpleBench | 25.1% | — |
| Chess Puzzles | 6% | — |
| LiveBench Reasoning | — | 15.9% |
| DTBench | 61.6% | — |
| LiveBench Data Analysis | — | 30.3% |
| LMCA | 9.8% | — |
| Adversarial NLI | — | 58.1% |
| BIG-Bench Hard | — | 79.1% |
| Epoch Capabilities Index | 127.25 | — |
| ForecastBench | 59.4 | — |
| HellaSwag | — | 77% |
| LiveBench | — | 24% |
| WinoGrande | — | 81.5% |
Math Phi 3 Small 8k Instruct leads
GPT-4 Turbo: 9.0 (#322), Phi 3 Small 8k Instruct: 27.6 (#248)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Math | 1272 | 1151 |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| OTIS Mock AIME 2024-2025 | 6.7% | — |
| LiveBench Math | — | 17.6% |
| MATH Level 5 | 46.7% | — |
Knowledge Phi 3 Small 8k Instruct leads
GPT-4 Turbo: 24.3 (#268), Phi 3 Small 8k Instruct: 29.1 (#240)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Expert | 1223 | 1067 |
| MMLU | 81.3% | 75.7% |
| GPQA Diamond | 46.6% | — |
| Confabulations | 28.4% | — |
| ARC (AI2) Challenge | — | 90.7% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 58.1% |
Multimodal Not comparable
GPT-4 Turbo: 30.6 (#110), Phi 3 Small 8k Instruct: —
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Vision | 1090 | — |
Multilingual GPT-4 Turbo leads
GPT-4 Turbo: 40.5 (#216), Phi 3 Small 8k Instruct: 28.5 (#272)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Non-English | 1245 | 1058 |
| LMArena Chinese | 1242 | 1061 |
| LMArena French | 1276 | 1135 |
| LMArena German | 1259 | 1080 |
| LMArena Japanese | 1194 | 966 |
| LMArena Korean | 1187 | 894 |
| LMArena Russian | 1259 | 1111 |
| LMArena Spanish | 1260 | 1111 |
Instruction Following GPT-4 Turbo leads
GPT-4 Turbo: 65.8 (#216), Phi 3 Small 8k Instruct: 51.9 (#292)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Instruction Following | 1249 | 1087 |
| LiveBench Instruction Following | — | 47.2% |
Long Context GPT-4 Turbo leads
GPT-4 Turbo: 38.0 (#206), Phi 3 Small 8k Instruct: 33.0 (#267)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Longer Query | 1254 | 1088 |
Writing & Preference GPT-4 Turbo leads
GPT-4 Turbo: 47.7 (#206), Phi 3 Small 8k Instruct: 31.1 (#284)
| Benchmark | GPT-4 Turbo | Phi 3 Small 8k Instruct |
|---|---|---|
| LMArena Text | 1272 | 1110 |
| LMArena Creative Writing | 1269 | 1083 |
| LMArena Multi-Turn | 1267 | 1068 |
| LiveBench Language | — | 12.9% |
Frequently asked questions
Is GPT-4 Turbo better than Phi 3 Small 8k Instruct?
GPT-4 Turbo is the stronger model overall, scoring 30.5 to 29.3 on the Noometry Index.
Is GPT-4 Turbo or Phi 3 Small 8k Instruct better for coding?
GPT-4 Turbo scores higher on coding benchmarks: 33.8 versus 27.9 in the Noometry coding category.
How many benchmarks do GPT-4 Turbo and Phi 3 Small 8k Instruct share?
18 benchmarks have published results for both models. GPT-4 Turbo has 36 scored results on Noometry and Phi 3 Small 8k Instruct has 32.