Model comparison
GPT-3.5-turbo vs o4-mini
o4-mini is the stronger model overall, scoring 41.6 to 23.1 on the Noometry Index. GPT-3.5-turbo costs 2.6× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. GPT-3.5-turbo scores higher in 0 categories and o4-mini in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where o4-mini leads 40.2 to 6.0.
- The biggest single-benchmark swing is MATH Level 5: 15.9% for GPT-3.5-turbo and 97.8% for o4-mini.
- GPT-3.5-turbo is cheaper at $0.50 / $1.50 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- o4-mini accepts more context: 200K tokens versus 16K.
Side by side
| GPT-3.5-turbo | o4-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 23.1 | 41.6 |
| Released | 2023-03-01 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 16K | 200K |
| Max output | 4K | 100K |
| Input $ / M tokens | $0.50 | $1.10 |
| Output $ / M tokens | $1.50 | $4.40 |
| Results tracked | 44 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
GPT-3.5-turbo: 23.5 (#326), o4-mini: 41.6 (#110)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| WeirdML | 3.5% | 52.6% |
| LMArena Coding | 1137 | 1369 |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| GSO | — | 3.6% |
| BigCodeBench Instruct | 39.1% | — |
| BigCodeBench Complete | 50.6% | — |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |
| HumanEval+ | 70.7% | — |
| MBPP+ | 69.7% | — |
Agentic & Tool Use Not comparable
GPT-3.5-turbo: —, o4-mini: 32.3 (#57)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| METR Time Horizons | 21.5% | 63.9% |
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
Reasoning o4-mini leads
GPT-3.5-turbo: 13.3 (#328), o4-mini: 24.6 (#143)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| Chess Puzzles | 0% | 26% |
| LMArena Hard Prompts | 1108 | 1351 |
| Mystery Game Puzzles | 3% | 5% |
| DTBench | 48.5% | 77.6% |
| LMCA | 9.7% | 26.5% |
| Epoch Capabilities Index | 118.48 | 145.63 |
| ForecastBench | 50.4 | 61.8 |
| ARC-AGI-2 | — | 6.1% |
| SimpleBench | — | 38.7% |
| Kagi LLM Benchmark | — | 67.6% |
| ARC-AGI-1 | — | 58.7% |
| CritPt | — | 0.6% |
| EnigmaEval | — | 9.2% |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| WinoGrande | 81.6% | — |
Math o4-mini leads
GPT-3.5-turbo: 6.0 (#321), o4-mini: 40.2 (#79)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0% | 36.1% |
| OTIS Mock AIME 2024-2025 | 2.2% | 81.7% |
| LMArena Math | 1142 | 1390 |
| MATH Level 5 | 15.9% | 97.8% |
| FrontierMath Tier 4 | — | 4.9% |
| Omni-MATH | — | 72% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
| GSM8K | 57.8% | — |
Knowledge o4-mini leads
GPT-3.5-turbo: 10.0 (#298), o4-mini: 43.5 (#83)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| GPQA Diamond | 28% | 79.6% |
| LMArena Expert | 1070 | 1344 |
| Humanity's Last Exam | — | 18.1% |
| SimpleQA Verified | — | 19.6% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
| ARC (AI2) Challenge | 87.4% | — |
| BoolQ | 87% | — |
| MMLU | 71.4% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |
Multimodal Not comparable
GPT-3.5-turbo: —, o4-mini: 39.5 (#46)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| LMArena Vision | — | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
Multilingual o4-mini leads
GPT-3.5-turbo: 31.6 (#253), o4-mini: 47.2 (#149)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| LMArena Non-English | 1108 | 1336 |
| LMArena Chinese | 1075 | 1355 |
| LMArena French | 1117 | 1362 |
| LMArena German | 1090 | 1336 |
| LMArena Japanese | 1043 | 1308 |
| LMArena Korean | 1019 | 1313 |
| LMArena Russian | 1123 | 1334 |
| LMArena Spanish | 1120 | 1346 |
Instruction Following o4-mini leads
GPT-3.5-turbo: 58.1 (#257), o4-mini: 75.4 (#63)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1119 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
GPT-3.5-turbo: 34.0 (#249), o4-mini: 45.5 (#31)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| LMArena Longer Query | 1121 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference o4-mini leads
GPT-3.5-turbo: 25.4 (#300), o4-mini: 54.2 (#148)
| Benchmark | GPT-3.5-turbo | o4-mini |
|---|---|---|
| LMArena Text | 1125 | 1353 |
| LMArena Creative Writing | 1092 | 1294 |
| LMArena Multi-Turn | 1117 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 451 | — |
| WildBench | — | 85.4% |
Frequently asked questions
Is GPT-3.5-turbo better than o4-mini?
o4-mini is the stronger model overall, scoring 41.6 to 23.1 on the Noometry Index. GPT-3.5-turbo costs 2.6× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Which is cheaper, GPT-3.5-turbo or o4-mini?
GPT-3.5-turbo is cheaper. It lists at $0.50 per million input tokens and $1.50 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is GPT-3.5-turbo or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 41.6 versus 23.5 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 16K.
How many benchmarks do GPT-3.5-turbo and o4-mini share?
29 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and o4-mini has 60.