Model comparison
GPT-4 vs GPT-5
GPT-5 is the stronger model overall, scoring 50.9 to 29.1 on the Noometry Index.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. GPT-4 scores higher in 0 categories and GPT-5 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 leads 55.0 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 91.4% for GPT-5.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $30 / $60 for GPT-4.
- GPT-5 accepts more context: 400K tokens versus 8K.
Side by side
| GPT-4 | GPT-5 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 29.1 | 50.9 |
| Released | 2023-03-14 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 8K | 400K |
| Max output | 8K | 128K |
| Input $ / M tokens | $30 | $1.25 |
| Output $ / M tokens | $60 | $10 |
| Results tracked | 38 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
GPT-4: 31.6 (#283), GPT-5: 50.3 (#47)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| WeirdML | 12.4% | 60.7% |
| LMArena Coding | 1254 | 1436 |
| SWE-bench Verified | — | 73.6% |
| SWE-bench Verified (bash only) | — | 65% |
| Aider Polyglot | — | 88% |
| LMArena WebDev | — | 1418 |
| SciCode | — | 42.9% |
| GSO | — | 6.9% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| ALE-Bench | — | 1,162 |
| AlgoTune | — | 1.67 |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, GPT-5: 33.1 (#56)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| METR Time Horizons | 36.1% | 69.6% |
| Terminal-Bench | — | 49.6% |
| GDPval | — | 34.8% |
| Remote Labor Index | — | 1.7% |
| DeepResearch Bench | — | 49.6% |
| BALROG | — | 32.8% |
| LMArena Search | — | 1133 |
Reasoning GPT-5 leads
GPT-4: 17.8 (#289), GPT-5: 38.3 (#64)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| Chess Puzzles | 4% | 37% |
| LMArena Hard Prompts | 1241 | 1416 |
| Mystery Game Puzzles | 12% | 23% |
| DTBench | 62.7% | 90.7% |
| LMCA | 17.1% | 40% |
| Epoch Capabilities Index | 125.89 | 150 |
| ForecastBench | 57.8 | 61.4 |
| ARC-AGI-2 | — | 9.9% |
| SimpleBench | — | 56.7% |
| Kagi LLM Benchmark | — | 72.7% |
| ARC-AGI-1 | — | 65.7% |
| CritPt | — | 12.6% |
| EnigmaEval | — | 10.5% |
| EBR-Bench | — | 12.7% |
| BIG-Bench Hard | 75.1% | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math GPT-5 leads
GPT-4: 10.8 (#309), GPT-5: 55.0 (#44)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 91.4% |
| LMArena Math | 1269 | 1407 |
| MATH Level 5 | 23% | 98.1% |
| FrontierMath (Tiers 1-3) | — | 55.4% |
| FrontierMath Tier 4 | — | 22% |
| ProofBench | — | 18% |
| Omni-MATH | — | 64.7% |
| FrontierMath (Feb 2025 set) | — | 32.4% |
| FrontierMath Tier 4 (v1) | — | 12.5% |
| GSM8K | 92% | — |
Knowledge GPT-5 leads
GPT-4: 18.4 (#282), GPT-5: 56.6 (#43)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| GPQA Diamond | 35.7% | 86.2% |
| LMArena Expert | 1211 | 1419 |
| Humanity's Last Exam | — | 25.3% |
| SimpleQA Verified | — | 50.1% |
| MMLU-Pro | — | 86.3% |
| Confabulations | — | 10.3% |
| Vectara Hallucination Rate | — | 14.7% |
| GPQA (HELM) | — | 79.2% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multimodal Not comparable
GPT-4: —, GPT-5: 46.8 (#13)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| LMArena Vision | — | 1232 |
| GeoBench | — | 81% |
| VPCT | — | 66% |
Multilingual GPT-5 leads
GPT-4: 40.6 (#215), GPT-5: 51.4 (#110)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| LMArena Non-English | 1246 | 1397 |
| LMArena Chinese | 1242 | 1422 |
| LMArena French | 1283 | 1410 |
| LMArena German | 1251 | 1416 |
| LMArena Japanese | 1209 | 1409 |
| LMArena Korean | 1184 | 1360 |
| LMArena Russian | 1251 | 1406 |
| LMArena Spanish | 1261 | 1399 |
Instruction Following GPT-5 leads
GPT-4: 65.3 (#222), GPT-5: 73.8 (#113)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| LMArena Instruction Following | 1241 | 1388 |
| IFEval | — | 87.5% |
Long Context GPT-5 leads
GPT-4: 37.7 (#212), GPT-5: 69.5 (#2)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| LMArena Longer Query | 1244 | 1399 |
| Fiction.LiveBench | — | 97.2% |
Writing & Preference GPT-5 leads
GPT-4: 34.9 (#268), GPT-5: 63.4 (#65)
| Benchmark | GPT-4 | GPT-5 |
|---|---|---|
| LMArena Text | 1263 | 1406 |
| LMArena Creative Writing | 1244 | 1365 |
| EQ-Bench Creative Writing | 752 | 1627 |
| LMArena Multi-Turn | 1257 | 1426 |
| Short-Story Creative Writing | — | 86% |
| WildBench | — | 85.7% |
Frequently asked questions
Is GPT-4 better than GPT-5?
GPT-5 is the stronger model overall, scoring 50.9 to 29.1 on the Noometry Index.
Which is cheaper, GPT-4 or GPT-5?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-4 lists at $30 and $60.
Is GPT-4 or GPT-5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 31.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5 does, with 400K tokens against 8K.
How many benchmarks do GPT-4 and GPT-5 share?
29 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and GPT-5 has 69.