Up to four models
Claude 3 Sonnet vs Claude 2.1
On the index (GPT-4 = 100), Claude 3 Sonnet scores 98 and Claude 2.1 96. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
Claude 3 Sonnet and Claude 2.1 are too close to call (2 points apart, within the uncertainty), ahead on 1 of 1 shared core tests.
Test by testthird-party results
Claude 3 SonnetClaude 2.1
Every result as a table
| Test | Claude 3 Sonnet | Claude 2.1 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 40.6% | 33.0% |
| MMLU (HELM runs) | 75.9% | 73.5% |
| Maths | ||
| AIME-style maths (OTIS mock) | 2.5% | 1.9% |
| Coding | ||
| WeirdML | 10.2% | 7.1% |
| Agents | ||
| DTBench | 53.6% | 50.9% |
Tests at least two of these models have taken. Bold rows are the core tests.