Up to four models
Claude 3 Opus vs Llama 2
On the index (GPT-4 = 100), Claude 3 Opus scores 108 and Llama 2 95. 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 Opus leads Llama 2 by 12 points, ahead on 1 of 1 shared core tests.
Test by testthird-party results
Claude 3 Opus
Llama 2
Every result as a table
| Test | Claude 3 Opus | Llama 2 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 47.2% | 26.3% |
| MMLU (HELM runs) | 84.6% | 69.5% |
| Maths | ||
| AIME-style maths (OTIS mock) | 4.7% | 0.0% |
| MATH (level 5) | 37.5% | 3.3% |
| Agents | ||
| DTBench | 61.6% | 41.6% |
Task length · METR18 sec
Tests at least two of these models have taken. Bold rows are the core tests.