R1 Distill Llama 70B
DeepSeek R1 Distill Llama 70B is a distilled large language model based on [Llama-3.3-70B-Instruct](/meta-llama/llama-3.3-70b-instruct), using outputs from [DeepSeek R1](/deepseek/deepseek-r1). The model leverages fine-tuning from DeepSeek R1's outputs, enabling competitive performance comparable to larger frontier models.
Input
Output
Context
131K
Max Output
131K
Parameters
70B
Technical Specifications
Model TypeChat
Context Window131K
Max Output Tokens131K
Parameters70B
Training Cutoff—
Licensemit
Capabilities
Input Modalities
text
Output Modalities
text
Features
code_generationfunction_callingjson_modereasoningstreamingtext_generation
Benchmarks
6.1%
HLE
67.0%
AIME
40.2%
GPQA
31.2%
SciCode
93.5%
MATH 500
79.5%
MMLU Pro
53.7
Math
11.4
Coding
26.6%
LiveCodeBench
16.0
Intelligence
45.7
Speed (tok/s)
44.1
TTFA (s)
0.389
TTFT (s)
Resources & Links
Estimated GPU Requirements for R1 Distill Llama 70B
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: May 5, 2026
Start building with the right model.
From model selection to production, one platform, no fragmentation.