Meta Llama 3.3 70B Instruct is a 70B multilingual instruction-tuned model using GQA, pretrained on 15T tokens with 25M+ synthetic fine-tuning examples and a 128K context window.
Capabilities, design details, and architectural traits
Llama 3.3 70B Instruct is Meta's instruction-tuned generative text model (text in/text out) positioned as the official successor to the instruction-tuned Llama 3.1 70B. Meta's official documentation states that developers should replace Llama 3.1 70B Instruct with this model wherever it was previously used.
Meta officially designates Llama 3.3 70B Instruct as superseding the instruction-tuned Llama 3.1 70B. The repository ships two versions: one for use with Hugging Face Transformers (requires version ≥ 4.45.0) and one for the original Llama codebase. A dynamic FP8 quantized variant is also available for latency-sensitive inference workloads.
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Meta | Anthropic | Anthropic |
| Release Date | December 6, 2024 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | Dec 2023 | May 2026 | - |
| Context & Limits | |||
| Context Window | 131K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.66 Best Input Pricing | $5 | $10 |
| Output Pricing | $0.72 Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | text | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 9.3 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | 11.9 | 78.0 Best Coding Index | 76.5 |
| Agentic Index | 0.3 | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.