Model comparison
Claude 2 vs Claude 3.5 Sonnet
Claude 3.5 Sonnet is the stronger model overall, scoring 34.6 to 25.0 on the Noometry Index.
Last verified . 7 shared benchmarks.
Summary
- They share 7 benchmarks with published results for both. Claude 2 scores higher in 0 categories and Claude 3.5 Sonnet in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Claude 3.5 Sonnet leads 28.6 to 16.9.
- The biggest single-benchmark swing is MATH Level 5: 11.7% for Claude 2 and 56.9% for Claude 3.5 Sonnet.
Side by side
| Claude 2 | Claude 3.5 Sonnet | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 25.0 | 34.6 |
| Released | 2023-07-11 | 2024-06-20 |
| Weights | Proprietary | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 8 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Claude 2: —, Claude 3.5 Sonnet: 39.0 (#165)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| HumanEval+ | 61.6% | 81.7% |
| Aider Polyglot | — | 51.6% |
| GSO | — | 4.6% |
| WeirdML | — | 40% |
| BigCodeBench Instruct | — | 46.8% |
| LiveBench Coding | — | 67.1% |
| LMArena Coding | — | 1342 |
| BigCodeBench Complete | — | 58.6% |
| CadEval | — | 48% |
| MBPP+ | — | 74.3% |
Agentic & Tool Use Not comparable
Claude 2: —, Claude 3.5 Sonnet: 32.3 (#67)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| TheAgentCompany | — | 24% |
| Cybench | — | 17.5% |
| BALROG | — | 32.6% |
| METR Time Horizons | — | 45.2% |
Reasoning Claude 3.5 Sonnet leads
Claude 2: 21.7 (#216), Claude 3.5 Sonnet: 23.1 (#183)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| DTBench | 51.9% | 67.8% |
| Epoch Capabilities Index | 120.13 | 133.55 |
| SimpleBench | — | 41.4% |
| EnigmaEval | — | 0.9% |
| LiveBench Reasoning | — | 56.7% |
| LMArena Hard Prompts | — | 1305 |
| LiveBench Data Analysis | — | 55% |
| ForecastBench | — | 60.7 |
| LiveBench | — | 59% |
Math Claude 3.5 Sonnet leads
Claude 2: 9.3 (#320), Claude 3.5 Sonnet: 19.2 (#288)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.5% | 8.5% |
| MATH Level 5 | 11.7% | 56.9% |
| Omni-MATH | — | 27.6% |
| LiveBench Math | — | 52.3% |
| LMArena Math | — | 1307 |
| FrontierMath (Feb 2025 set) | — | 2.1% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Claude 3.5 Sonnet leads
Claude 2: 16.9 (#287), Claude 3.5 Sonnet: 28.6 (#245)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| GPQA Diamond | 34.7% | 55.3% |
| MMLU | 78.5% | 87.3% |
| Humanity's Last Exam | — | 4.1% |
| MMLU-Pro | — | 77.7% |
| Confabulations | — | 19.9% |
| GPQA (HELM) | — | 56.5% |
| LMArena Expert | — | 1265 |
| TriviaQA | 87.5% | — |
Multimodal Not comparable
Claude 2: —, Claude 3.5 Sonnet: 26.5 (#120)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| LMArena Vision | — | 1125 |
| Video-MME | — | 60% |
| GeoBench | — | 62% |
| VPCT | — | 33% |
Multilingual Not comparable
Claude 2: —, Claude 3.5 Sonnet: 43.2 (#185)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| LMArena Non-English | — | 1283 |
| LMArena Chinese | — | 1272 |
| LMArena French | — | 1305 |
| LMArena German | — | 1297 |
| LMArena Japanese | — | 1234 |
| LMArena Korean | — | 1200 |
| LMArena Russian | — | 1306 |
| LMArena Spanish | — | 1290 |
Instruction Following Not comparable
Claude 2: —, Claude 3.5 Sonnet: 68.8 (#182)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| LiveBench Instruction Following | — | 69.3% |
| IFEval | — | 85.5% |
| LMArena Instruction Following | — | 1297 |
Long Context Not comparable
Claude 2: —, Claude 3.5 Sonnet: 39.9 (#167)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| LMArena Longer Query | — | 1311 |
Writing & Preference Not comparable
Claude 2: —, Claude 3.5 Sonnet: 52.9 (#164)
| Benchmark | Claude 2 | Claude 3.5 Sonnet |
|---|---|---|
| LMArena Text | — | 1298 |
| LMArena Creative Writing | — | 1292 |
| Short-Story Creative Writing | — | 80.3% |
| EQ-Bench Creative Writing | — | 1451 |
| WildBench | — | 79.2% |
| LMArena Multi-Turn | — | 1326 |
| LiveBench Language | — | 53.8% |
Frequently asked questions
Is Claude 2 better than Claude 3.5 Sonnet?
Claude 3.5 Sonnet is the stronger model overall, scoring 34.6 to 25.0 on the Noometry Index.
How many benchmarks do Claude 2 and Claude 3.5 Sonnet share?
7 benchmarks have published results for both models. Claude 2 has 8 scored results on Noometry and Claude 3.5 Sonnet has 60.