Llama 3.3 70B Instruct scores 8.6 on the Respan Index (#137), ahead of Phi-4 at 7.0 (#138). Phi-4 leads on 4 of the 6 benchmarks both report. Llama 3.3 70B Instruct has the larger context window (131K tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 68.9% |
| 70.4% |
| BALROG (BALROG) | 23% | 11.6% |
Higher is better. Each score links to where it was published.