Llama 3.1 8B Instruct
Zero-eval
#1MBPP EvalPlus (base)
#2GSM-8K (CoT)
#2MATH (CoT)
+8 more
by Meta
+
+
+
+
About
Llama 3.1 8B Instruct is a language model developed by Meta. It achieves strong performance with an average score of 61.3% across 18 benchmarks. It excels particularly in GSM-8K (CoT) (84.5%), ARC-C (83.4%), API-Bank (82.6%). It supports a 262K token context window for handling large documents. The model is available through 9 API providers. Released in 2024, it represents Meta's latest advancement in AI technology.
+
+
+
+
Pricing Range
Input (per 1M)$0.03 -$0.22
Output (per 1M)$0.03 -$0.22
Providers9
+
+
+
+
Timeline
AnnouncedJul 23, 2024
ReleasedJul 23, 2024
Knowledge CutoffDec 31, 2023
+
+
+
+
Specifications
Training Tokens15.0T
+
+
+
+
License & Family
License
Llama 3.1 Community License
Performance Overview
Performance metrics and category breakdown
Overall Performance
18 benchmarks
Average Score
61.3%
Best Score
84.5%
High Performers (80%+)
4Performance Metrics
Max Context Window
262.1KAvg Throughput
532.6 tok/sAvg Latency
0ms+
+
+
+
All Benchmark Results for Llama 3.1 8B Instruct
Complete list of benchmark scores with detailed information
| GSM-8K (CoT) | text | 0.84 | 84.5% | Self-reported | |
| ARC-C | text | 0.83 | 83.4% | Self-reported | |
| API-Bank | text | 0.83 | 82.6% | Self-reported | |
| IFEval | text | 0.80 | 80.4% | Self-reported | |
| BFCL | text | 0.76 | 76.1% | Self-reported | |
| MMLU (CoT) | text | 0.73 | 73.0% | Self-reported | |
| MBPP EvalPlus (base) | text | 0.73 | 72.8% | Self-reported | |
| HumanEval | text | 0.73 | 72.6% | Self-reported | |
| MMLU | text | 0.69 | 69.4% | Self-reported | |
| Multilingual MGSM (CoT) | text | 0.69 | 68.9% | Self-reported |
Showing 1 to 10 of 18 benchmarks