← All benchmarks
Llama 3.3 Instruct 70B
Meta · Released 2024-12-06 · Independent benchmark scores from Artificial Analysis
Intelligence Index
7.7
Output speed
94.2
tokens/sec
Blended price
$0.71
per 1M tokens
Time to first token
0.69s
Model metadata
| Model ID | 976cc8ad-7904-4056-83c5-960181f47d5f |
|---|---|
| Slug | llama-3-3-instruct-70b |
| Release date | 2024-12-06 |
| Provider | Meta |
| Provider slug | meta |
| Output speed | 94.2 tok/s |
| Time to first token | 0.69s |
| Time to first answer | 0.69s |
| Blended price ($/1M) | $0.71 |
| Input price ($/1M) | $0.71 |
| Output price ($/1M) | $0.72 |
All evaluation scores (18)
| aime | 30.0% |
|---|---|
| aime 25 | 7.7% |
| artificial analysis coding index | 11.90 |
| artificial analysis intelligence index | 7.70 |
| artificial analysis math index | 7.70 |
| gpqa | 49.8% |
| hle | 3.6% |
| ifbench | 47.1% |
| lcr | 15.7% |
| livecodebench | 28.8% |
| math 500 | 77.3% |
| mmlu pro | 71.3% |
| scicode | — |
| tau banking | — |
| tau2 | 26.6% |
| terminalbench hard | 3.0% |
| terminalbench v2 1 | 4.9% |
| terminalbench v4 0 | — |
Source: artificialanalysis.ai/models/llama-3-3-instruct-70b
Benchmark data from Artificial Analysis.