Capability
Code Generation Across Programming Languages
20 artifacts provide this capability.
Want a personalized recommendation?
Find the best match →Top Matches
1
CodeLlama 70BModel47/100
via “multi-language code generation from natural language prompts”
Meta's 70B specialized code generation model.
Unique: Largest open-source dedicated code model (70B parameters) trained on 1 trillion code tokens with explicit multi-language support across 15+ languages, compared to general-purpose LLMs fine-tuned on mixed data. Specialized variants (Python-only, instruction-tuned) allow task-specific optimization without retraining.
vs others: Outperforms smaller open-source code models (CodeGen, PolyCoder) on HumanEval and supports more languages than GPT-3.5-Codex while remaining fully open-source and commercially usable without API dependencies.