--- title: "human-eval vs LLMs-from-scratch" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/openai-human-eval-vs-rasbt-llms-from-scratch" tools: ["openai-human-eval", "rasbt-llms-from-scratch"] --- # human-eval vs LLMs-from-scratch *GraphCanon updated Jul 11, 2026* ## Verdict Pick human-eval when human-eval is primarily Python; LLMs-from-scratch is Jupyter Notebook; pick LLMs-from-scratch when lLMs-from-scratch is primarily Jupyter Notebook; human-eval is Python. [human-eval](https://github.com/openai/human-eval) reports 3.3k GitHub stars, 449 forks, and 42 open issues, last pushed Jan 17, 2025. [LLMs-from-scratch](https://amzn.to/4fqvn0D) has 99k stars, 15k forks, and 4 open issues, last pushed Jun 2, 2026. Figures are from public GitHub metadata via [human-eval's repository](https://github.com/openai/human-eval) and [LLMs-from-scratch's repository](https://github.com/rasbt/LLMs-from-scratch). | | [human-eval](/tools/openai-human-eval.md) | [LLMs-from-scratch](/tools/rasbt-llms-from-scratch.md) | | --- | --- | --- | | Tagline | Code for the paper "Evaluating Large Language Models Trained on Code" | Implement a ChatGPT-like LLM in PyTorch from scratch, step by step | | Stars | 3,294 | 98,899 | | Forks | 449 | 15,183 | | Open issues | 42 | 4 | | Language | Python | Jupyter Notebook | | Adopt for | - | LLMs-from-scratch is a project-oriented repository aimed at building PyTorch-based language models from the ground up, with detailed step-by-step instructions. | | Persona | - | - | | Runtime | - | - | | License | MIT | Other | | Categories | Model Training, LLM Frameworks, Evaluation & Observability | LLM Frameworks, Model Training | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [human-eval](/tools/openai-human-eval.md) | [LLMs-from-scratch](/tools/rasbt-llms-from-scratch.md) | | --- | --- | --- | | Maintenance | Dormant (18%) | Steady (60%) | | Days since push | 540d | 38d | | Open issues (now) | 42 | 4 | | Owner type | Organization | User | | Security scan | No criticals | No lockfile | | Full report | [trust report](/tools/openai-human-eval/trust.md) | [trust report](/tools/rasbt-llms-from-scratch/trust.md) | ## Decision facts: LLMs-from-scratch - **Adopt for:** LLMs-from-scratch is a project-oriented repository aimed at building PyTorch-based language models from the ground up, with detailed step-by-step instructions. ## Choose when ### Choose human-eval if… - human-eval is primarily Python; LLMs-from-scratch is Jupyter Notebook. - License: human-eval is MIT, LLMs-from-scratch is Other. - Tags unique to human-eval: python. - Also covers Evaluation & Observability. ### Choose LLMs-from-scratch if… - LLMs-from-scratch is primarily Jupyter Notebook; human-eval is Python. - License: LLMs-from-scratch is Other, human-eval is MIT. - Tags unique to LLMs-from-scratch: deep-learning, ai, artificial-intelligence, attention-mechanism. - - You are an advanced practitioner aiming to fully understand the underpinnings of LLMs using PyTorch as your primary framework. ## When NOT to use human-eval - Last GitHub push was 540 days ago (dormant maintenance, Jan 17, 2025). Validate activity before betting a new project on human-eval. - Model Training: Try prompting and RAG first; fine-tuning is the answer to style/format, not missing knowledge. - LLM Frameworks: Avoid a framework for a single prompt-and-retrieve call; the abstraction can cost more than it saves. - Evaluation & Observability: Defer heavyweight eval infra only until you have real traffic - never skip it once users depend on answers. ## When NOT to use LLMs-from-scratch - - If you are looking for a rapid deployment of an LLM without understanding its intricate structure - this tool requires extensive manual and conceptual work. - - You prefer frameworks with automatic model generation or other high-level abstractions that simplify the process. This repository emphasizes manual creation, which is more time-consuming but offers丰 - a deeper learning experience. ## Common questions ### What is the difference between human-eval and LLMs-from-scratch? human-eval: Code for the paper "Evaluating Large Language Models Trained on Code". LLMs-from-scratch: Implement a ChatGPT-like LLM in PyTorch from scratch, step by step. See the comparison table for live GitHub stats and shared categories. ### When should I choose human-eval over LLMs-from-scratch? Choose human-eval over LLMs-from-scratch when human-eval is primarily Python; LLMs-from-scratch is Jupyter Notebook; License: human-eval is MIT, LLMs-from-scratch is Other; Tags unique to human-eval: python; Also covers Evaluation & Observability. ### When should I choose LLMs-from-scratch over human-eval? Choose LLMs-from-scratch over human-eval when LLMs-from-scratch is primarily Jupyter Notebook; human-eval is Python; License: LLMs-from-scratch is Other, human-eval is MIT; Tags unique to LLMs-from-scratch: deep-learning, ai, artificial-intelligence, attention-mechanism; - You are an advanced practitioner aiming to fully understand the underpinnings of LLMs using PyTorch as your primary framework. ### When should I avoid human-eval? Last GitHub push was 540 days ago (dormant maintenance, Jan 17, 2025). Validate activity before betting a new project on human-eval. Model Training: Try prompting and RAG first; fine-tuning is the answer to style/format, not missing knowledge. LLM Frameworks: Avoid a framework for a single prompt-and-retrieve call; the abstraction can cost more than it saves. Evaluation & Observability: Defer heavyweight eval infra only until you have real traffic - never skip it once users depend on answers. ### When should I avoid LLMs-from-scratch? - If you are looking for a rapid deployment of an LLM without understanding its intricate structure - this tool requires extensive manual and conceptual work. - You prefer frameworks with automatic model generation or other high-level abstractions that simplify the process. This repository emphasizes manual creation, which is more time-consuming but offers丰 a deeper learning experience. ### Is human-eval or LLMs-from-scratch more popular on GitHub? LLMs-from-scratch has more GitHub stars (98,899 vs 3,294). Stars measure visibility, not whether either tool fits your constraints. ### Are human-eval and LLMs-from-scratch open source? Yes - both are open-source projects on GitHub (human-eval: MIT, LLMs-from-scratch: Other). ### Where can I find alternatives to human-eval or LLMs-from-scratch? GraphCanon lists graph-backed alternatives at [human-eval alternatives](/tools/openai-human-eval/alternatives) and [LLMs-from-scratch alternatives](/tools/rasbt-llms-from-scratch/alternatives) ([human-eval markdown twin](/tools/openai-human-eval/alternatives.md), [LLMs-from-scratch markdown twin](/tools/rasbt-llms-from-scratch/alternatives.md)), ranked by typed relationship edges rather than popularity votes. ### Is there a machine-readable version of this comparison? Yes. The markdown twin at [this comparison](/compare/openai-human-eval-vs-rasbt-llms-from-scratch.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, human-eval or LLMs-from-scratch? human-eval: Dormant. LLMs-from-scratch: Steady. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance. ### Where are the full trust reports for human-eval and LLMs-from-scratch? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [human-eval trust report](/tools/openai-human-eval/trust); [LLMs-from-scratch trust report](/tools/rasbt-llms-from-scratch/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=openai-human-eval`](/api/graphcanon/graph?tool=openai-human-eval) - LLM index: [/llms.txt](/llms.txt) - Full corpus: [/llms-full.txt](/llms-full.txt) _GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_