Model comparison
GPT-3.5-turbo vs GPT-5.2
GPT-5.2 is the stronger model overall, scoring 54.1 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 6.4× less per token, which makes it the better buy when GPT-5.2'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 GPT-5.2 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.2 leads 60.0 to 6.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.2% for GPT-3.5-turbo and 96.1% for GPT-5.2.
- GPT-3.5-turbo is cheaper at $0.50 / $1.50 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
- GPT-5.2 accepts more context: 400K tokens versus 16K.
Side by side
| GPT-3.5-turbo | GPT-5.2 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 23.2 | 54.1 |
| Released | 2023-03-01 | 2025-12-11 |
| Weights | Proprietary | Proprietary |
| Context window | 16K | 400K |
| Max output | 4K | 128K |
| Input $ / M tokens | $0.50 | $1.75 |
| Output $ / M tokens | $1.50 | $14 |
| Results tracked | 44 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.2 leads
GPT-3.5-turbo: 23.9 (#331), GPT-5.2: 51.6 (#37)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| WeirdML | 3.5% | 72.2% |
| LMArena Coding | 1136 | 1447 |
| SWE-bench Verified | — | 73.8% |
| SWE-bench Verified (bash only) | — | 72.8% |
| LMArena WebDev | — | 1416 |
| SWE-bench Multilingual | — | 66.7% |
| GSO | — | 27.4% |
| BigCodeBench Instruct | 39.1% | — |
| BigCodeBench Complete | 50.6% | — |
| ALE-Bench | — | 1,294 |
| AlgoTune | — | 2.05 |
| HumanEval+ | 70.7% | — |
| MBPP+ | 69.7% | — |
Agentic & Tool Use Not comparable
GPT-3.5-turbo: —, GPT-5.2: 40.2 (#24)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| METR Time Horizons | 21.5% | 75.3% |
| Terminal-Bench | — | 64.9% |
| Berkeley Function Calling Leaderboard | — | 55.9% |
| GDPval | — | 49.7% |
| Remote Labor Index | — | 2.5% |
| τ²-bench Airline | — | 83% |
| τ²-bench Banking | — | 32.2% |
| τ²-bench Retail | — | 81.6% |
| τ²-bench Telecom | — | 89.7% |
| DeepResearch Bench | — | 41.1% |
| LMArena Search | — | 1207 |
| Vending-Bench 2 | — | 3,591 |
Reasoning GPT-5.2 leads
GPT-3.5-turbo: 13.8 (#332), GPT-5.2: 50.2 (#35)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| Chess Puzzles | 0% | 49% |
| LMArena Hard Prompts | 1108 | 1445 |
| Mystery Game Puzzles | 3% | 23% |
| DTBench | 48.5% | 90.9% |
| LMCA | 9.7% | 43.9% |
| Epoch Capabilities Index | 118.55 | 153.45 |
| ForecastBench | 50.4 | 60.1 |
| ARC-AGI-2 | — | 52.9% |
| SimpleBench | — | 45.8% |
| Kagi LLM Benchmark | — | 73.3% |
| NYT Connections (extended) | — | 83.6% |
| ARC-AGI-1 | — | 86.2% |
| EnigmaEval | — | 10.4% |
| EBR-Bench | — | 23% |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| WinoGrande | 81.6% | — |
Math GPT-5.2 leads
GPT-3.5-turbo: 6.3 (#327), GPT-5.2: 60.0 (#38)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0% | 67.4% |
| OTIS Mock AIME 2024-2025 | 2.2% | 96.1% |
| LMArena Math | 1142 | 1440 |
| FrontierMath Tier 4 | — | 31.7% |
| MathArena Final-Answer Competitions | — | 72% |
| ProofBench | — | 15% |
| MATH Level 5 | 15.9% | — |
| FrontierMath (Feb 2025 set) | — | 40.7% |
| FrontierMath Tier 4 (v1) | — | 18.8% |
| GSM8K | 57.8% | — |
Knowledge GPT-5.2 leads
GPT-3.5-turbo: 10.0 (#303), GPT-5.2: 59.3 (#32)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| GPQA Diamond | 28% | 91.4% |
| LMArena Expert | 1070 | 1445 |
| Humanity's Last Exam | — | 27.8% |
| SimpleQA Verified | — | 37.1% |
| Vectara Hallucination Rate | — | 8.4% |
| ARC (AI2) Challenge | 87.4% | — |
| BoolQ | 87% | — |
| MMLU | 71.4% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |
Multimodal Not comparable
GPT-3.5-turbo: —, GPT-5.2: 51.3 (#7)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| LMArena Vision | — | 1268 |
| VPCT | — | 84% |
| Furniture Assembly | — | 38.3% |
| LMArena Document | — | 1405 |
Multilingual GPT-5.2 leads
GPT-3.5-turbo: 31.5 (#258), GPT-5.2: 53.4 (#67)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| LMArena Non-English | 1108 | 1425 |
| LMArena Chinese | 1075 | 1460 |
| LMArena French | 1118 | 1455 |
| LMArena German | 1090 | 1448 |
| LMArena Japanese | 1043 | 1420 |
| LMArena Korean | 1019 | 1392 |
| LMArena Russian | 1123 | 1440 |
| LMArena Spanish | 1121 | 1433 |
Instruction Following GPT-5.2 leads
GPT-3.5-turbo: 57.9 (#262), GPT-5.2: 74.7 (#89)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| LMArena Instruction Following | 1119 | 1417 |
Long Context GPT-5.2 leads
GPT-3.5-turbo: 34.0 (#254), GPT-5.2: 44.0 (#78)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| LMArena Longer Query | 1121 | 1428 |
| CL-bench | — | 18.2% |
Writing & Preference GPT-5.2 leads
GPT-3.5-turbo: 25.3 (#305), GPT-5.2: 66.8 (#32)
| Benchmark | GPT-3.5-turbo | GPT-5.2 |
|---|---|---|
| LMArena Text | 1125 | 1439 |
| LMArena Creative Writing | 1092 | 1401 |
| EQ-Bench Creative Writing | 451 | 1703 |
| LMArena Multi-Turn | 1117 | 1458 |
Frequently asked questions
Is GPT-3.5-turbo better than GPT-5.2?
GPT-5.2 is the stronger model overall, scoring 54.1 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 6.4× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Which is cheaper, GPT-3.5-turbo or GPT-5.2?
GPT-3.5-turbo is cheaper. It lists at $0.50 per million input tokens and $1.50 per million output tokens; GPT-5.2 lists at $1.75 and $14.
Is GPT-3.5-turbo or GPT-5.2 better for coding?
GPT-5.2 scores higher on coding benchmarks: 51.6 versus 23.9 in the Noometry coding category.
Which has the bigger context window?
GPT-5.2 does, with 400K tokens against 16K.
How many benchmarks do GPT-3.5-turbo and GPT-5.2 share?
29 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and GPT-5.2 has 67.