Model comparison
Claude Sonnet 4 vs GPT-3.5-turbo
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 8.0× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Last verified . 27 shared benchmarks.
Summary
- They share 27 benchmarks with published results for both. Claude Sonnet 4 scores higher in 7 categories and GPT-3.5-turbo in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 4 leads 43.3 to 6.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 71.1% for Claude Sonnet 4 and 2.2% for GPT-3.5-turbo.
- GPT-3.5-turbo is cheaper at $0.50 / $1.50 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Claude Sonnet 4 accepts more context: 200K tokens versus 16K.
Side by side
| Claude Sonnet 4 | GPT-3.5-turbo | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 40.8 | 23.2 |
| Released | 2025-05-22 | 2023-03-01 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 16K |
| Max output | 64K | 4K |
| Input $ / M tokens | $3 | $0.50 |
| Output $ / M tokens | $15 | $1.50 |
| Results tracked | 58 | 44 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 43.5 (#88), GPT-3.5-turbo: 23.9 (#331)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| WeirdML | 46.1% | 3.5% |
| LMArena Coding | 1414 | 1136 |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| SciCode | 40% | — |
| GSO | 4.9% | — |
| BigCodeBench Instruct | — | 39.1% |
| BigCodeBench Complete | — | 50.6% |
| ALE-Bench | 655.35 | — |
| HumanEval+ | — | 70.7% |
| MBPP+ | — | 69.7% |
Agentic & Tool Use Not comparable
Claude Sonnet 4: 38.5 (#31), GPT-3.5-turbo: —
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| METR Time Horizons | 62% | 21.5% |
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
Reasoning Claude Sonnet 4 leads
Claude Sonnet 4: 22.9 (#187), GPT-3.5-turbo: 13.8 (#332)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Hard Prompts | 1372 | 1108 |
| DTBench | 77.1% | 48.5% |
| LMCA | 29% | 9.7% |
| Epoch Capabilities Index | 141.69 | 118.55 |
| ForecastBench | 60.2 | 50.4 |
| ARC-AGI-2 | 5.9% | — |
| SimpleBench | 45.5% | — |
| Kagi LLM Benchmark | 73% | — |
| ARC-AGI-1 | 40% | — |
| CritPt | 0.3% | — |
| Chess Puzzles | — | 0% |
| EnigmaEval | 3.1% | — |
| Mystery Game Puzzles | — | 3% |
| Adversarial NLI | — | 58.1% |
| BIG-Bench Hard | — | 61.6% |
| CommonsenseQA 2.0 | — | 57% |
| WinoGrande | — | 81.6% |
Math Claude Sonnet 4 leads
Claude Sonnet 4: 43.3 (#80), GPT-3.5-turbo: 6.3 (#327)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 2.2% |
| LMArena Math | 1375 | 1142 |
| MATH Level 5 | 84.4% | 15.9% |
| FrontierMath (Tiers 1-3) | — | 0% |
| Omni-MATH | 60.2% | — |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
| GSM8K | — | 57.8% |
Knowledge Claude Sonnet 4 leads
Claude Sonnet 4: 41.8 (#108), GPT-3.5-turbo: 10.0 (#303)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| GPQA Diamond | 79.2% | 28% |
| LMArena Expert | 1372 | 1070 |
| Humanity's Last Exam | 7.8% | — |
| MMLU-Pro | 84.3% | — |
| Confabulations | 13.2% | — |
| Vectara Hallucination Rate | 10.3% | — |
| GPQA (HELM) | 70.6% | — |
| ARC (AI2) Challenge | — | 87.4% |
| BoolQ | — | 87% |
| MMLU | — | 71.4% |
| OpenBookQA | — | 86% |
| TriviaQA | — | 85.8% |
Multimodal Not comparable
Claude Sonnet 4: 26.2 (#121), GPT-3.5-turbo: —
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |
Multilingual Claude Sonnet 4 leads
Claude Sonnet 4: 46.7 (#156), GPT-3.5-turbo: 31.5 (#258)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Non-English | 1333 | 1108 |
| LMArena Chinese | 1350 | 1075 |
| LMArena French | 1363 | 1118 |
| LMArena German | 1331 | 1090 |
| LMArena Japanese | 1302 | 1043 |
| LMArena Korean | 1291 | 1019 |
| LMArena Russian | 1355 | 1123 |
| LMArena Spanish | 1357 | 1121 |
Instruction Following Claude Sonnet 4 leads
Claude Sonnet 4: 71.7 (#145), GPT-3.5-turbo: 57.9 (#262)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Instruction Following | 1376 | 1119 |
| IFEval | 84% | — |
Long Context Too close to call
Claude Sonnet 4: 33.7 (#259), GPT-3.5-turbo: 34.0 (#254)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Longer Query | 1398 | 1121 |
| Fiction.LiveBench | 46.9% | — |
Writing & Preference Claude Sonnet 4 leads
Claude Sonnet 4: 57.1 (#132), GPT-3.5-turbo: 25.3 (#305)
| Benchmark | Claude Sonnet 4 | GPT-3.5-turbo |
|---|---|---|
| LMArena Text | 1351 | 1125 |
| LMArena Creative Writing | 1345 | 1092 |
| EQ-Bench Creative Writing | 1483 | 451 |
| LMArena Multi-Turn | 1376 | 1117 |
| Short-Story Creative Writing | 81.4% | — |
| WildBench | 83.8% | — |
Frequently asked questions
Is Claude Sonnet 4 better than GPT-3.5-turbo?
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 8.0× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4 or GPT-3.5-turbo?
GPT-3.5-turbo is cheaper. It lists at $0.50 per million input tokens and $1.50 per million output tokens; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or GPT-3.5-turbo better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 43.5 versus 23.9 in the Noometry coding category.
Which has the bigger context window?
Claude Sonnet 4 does, with 200K tokens against 16K.
How many benchmarks do Claude Sonnet 4 and GPT-3.5-turbo share?
27 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and GPT-3.5-turbo has 44.