Model comparison
Claude 3.5 Sonnet vs GPT-5.4
GPT-5.4 is the stronger model overall, scoring 59.5 to 34.4 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Claude 3.5 Sonnet scores higher in 0 categories and GPT-5.4 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 leads 73.3 to 18.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 8.5% for Claude 3.5 Sonnet and 97.8% for GPT-5.4.
Side by side
| Claude 3.5 Sonnet | GPT-5.4 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 34.4 | 59.5 |
| Released | 2024-06-20 | 2026-03-05 |
| Weights | Proprietary | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 128K |
| Input $ / M tokens | — | $2.50 |
| Output $ / M tokens | — | $15 |
| Results tracked | 60 | 68 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 leads
Claude 3.5 Sonnet: 39.5 (#145), GPT-5.4: 52.8 (#28)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| GSO | 4.6% | 31.4% |
| WeirdML | 40% | 77.7% |
| LMArena Coding | 1343 | 1497 |
| SWE-bench Verified | — | 76.9% |
| DeepSWE | — | 51.8% |
| Aider Polyglot | 51.6% | — |
| LMArena WebDev | — | 1465 |
| SciCode | — | 56.6% |
| BigCodeBench Instruct | 46.8% | — |
| LiveBench Coding | 67.1% | — |
| MirrorCode | — | 15.6% |
| BigCodeBench Complete | 58.6% | — |
| CadEval | 48% | — |
| ALE-Bench | — | 1,607 |
| AlgoTune | — | 1.85 |
| HumanEval+ | 81.7% | — |
| MBPP+ | 74.3% | — |
Agentic & Tool Use GPT-5.4 leads
Claude 3.5 Sonnet: 32.1 (#61), GPT-5.4: 46.5 (#12)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| METR Time Horizons | 45.2% | 74.3% |
| Terminal-Bench | — | 81.8% |
| APEX-Agents | — | 52.4% |
| TheAgentCompany | 24% | — |
| τ²-bench Banking | — | 39.4% |
| Cybench | 17.5% | — |
| DeepResearch Bench | — | 35.1% |
| PostTrainBench | — | 19% |
| BALROG | 32.6% | — |
| GBAEval | — | 45.1% |
| LMArena Search | — | 1197 |
| Vending-Bench 2 | — | 6,144 |
Reasoning GPT-5.4 leads
Claude 3.5 Sonnet: 22.5 (#176), GPT-5.4: 61.7 (#17)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| EnigmaEval | 0.9% | 16% |
| LMArena Hard Prompts | 1305 | 1485 |
| DTBench | 67.8% | 94.4% |
| Epoch Capabilities Index | 133.55 | 156.92 |
| ForecastBench | 60.7 | 59.5 |
| ARC-AGI-2 | — | 74% |
| SimpleBench | 41.4% | — |
| Kagi LLM Benchmark | — | 63.8% |
| NYT Connections (extended) | — | 91.3% |
| ARC-AGI-1 | — | 93.7% |
| CritPt | — | 23.4% |
| Chess Puzzles | — | 44% |
| Thematic Generalization | — | 80% |
| EBR-Bench | — | 25.4% |
| LiveBench Reasoning | 56.7% | — |
| Mystery Game Puzzles | — | 37% |
| LiveBench Data Analysis | 55% | — |
| LMCA | — | 52% |
| LiveBench | 59% | — |
Math GPT-5.4 leads
Claude 3.5 Sonnet: 18.0 (#282), GPT-5.4: 73.3 (#14)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 8.5% | 97.8% |
| LMArena Math | 1306 | 1488 |
| FrontierMath (Feb 2025 set) | 2.1% | 47.6% |
| FrontierMath Tier 4 (v1) | 0% | 27.1% |
| FrontierMath (Tiers 1-3) | — | 78.6% |
| FrontierMath Tier 4 | — | 49% |
| MathArena Final-Answer Competitions | — | 83.1% |
| ProofBench | — | 56% |
| Omni-MATH | 27.6% | — |
| LiveBench Math | 52.3% | — |
| MATH Level 5 | 56.9% | — |
Knowledge GPT-5.4 leads
Claude 3.5 Sonnet: 28.5 (#236), GPT-5.4: 65.3 (#13)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| GPQA Diamond | 55.3% | 93.3% |
| Humanity's Last Exam | 4.1% | 36.2% |
| LMArena Expert | 1264 | 1507 |
| SimpleQA Verified | — | 45.1% |
| MMLU-Pro | 77.7% | — |
| Confabulations | 19.9% | — |
| Vectara Hallucination Rate | — | 7% |
| GPQA (HELM) | 56.5% | — |
| MMLU | 87.3% | — |
Multimodal GPT-5.4 leads
Claude 3.5 Sonnet: 26.0 (#115), GPT-5.4: 43.8 (#19)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Vision | 1125 | 1302 |
| Video-MME | 60% | — |
| GeoBench | 62% | — |
| VPCT | 33% | — |
| Blueprint-Bench 2 | — | 27.1% |
| Furniture Assembly | — | 37.5% |
| LMArena Document | — | 1471 |
Multilingual GPT-5.4 leads
Claude 3.5 Sonnet: 43.4 (#180), GPT-5.4: 56.4 (#22)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Non-English | 1283 | 1465 |
| LMArena Chinese | 1271 | 1520 |
| LMArena French | 1304 | 1493 |
| LMArena German | 1297 | 1472 |
| LMArena Japanese | 1234 | 1484 |
| LMArena Korean | 1200 | 1449 |
| LMArena Russian | 1305 | 1480 |
| LMArena Spanish | 1289 | 1453 |
Instruction Following GPT-5.4 leads
Claude 3.5 Sonnet: 69.0 (#177), GPT-5.4: 77.3 (#25)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Instruction Following | 1297 | 1469 |
| LiveBench Instruction Following | 69.3% | — |
| IFEval | 85.5% | — |
Long Context GPT-5.4 leads
Claude 3.5 Sonnet: 39.9 (#162), GPT-5.4: 50.3 (#8)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Longer Query | 1311 | 1474 |
| CL-bench | — | 27.9% |
| CL-bench Life | — | 21.7% |
Writing & Preference GPT-5.4 leads
Claude 3.5 Sonnet: 53.0 (#159), GPT-5.4: 72.1 (#18)
| Benchmark | Claude 3.5 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Text | 1298 | 1469 |
| LMArena Creative Writing | 1292 | 1439 |
| EQ-Bench Creative Writing | 1451 | 1840 |
| LMArena Multi-Turn | 1326 | 1482 |
| Short-Story Creative Writing | 80.3% | — |
| WildBench | 79.2% | — |
| EQ-Bench 4 | — | 1272 |
| LiveBench Language | 53.8% | — |
Frequently asked questions
Is Claude 3.5 Sonnet better than GPT-5.4?
GPT-5.4 is the stronger model overall, scoring 59.5 to 34.4 on the Noometry Index.
Is Claude 3.5 Sonnet or GPT-5.4 better for coding?
GPT-5.4 scores higher on coding benchmarks: 52.8 versus 39.5 in the Noometry coding category.
How many benchmarks do Claude 3.5 Sonnet and GPT-5.4 share?
31 benchmarks have published results for both models. Claude 3.5 Sonnet has 60 scored results on Noometry and GPT-5.4 has 68.