Stanford CS336: Language Modeling from Scratch: study plan and how long it takes
The Stanford Online playlist is 17 videos and 22.3 hours long. Here's how long it takes at different paces, a week-by-week schedule built from the real videos, and a planner that turns it into your own daily plan.
17 videos · 22.3 hours · Stanford Online
Stanford CS336 (2025), taught by Percy Liang and Tatsunori Hashimoto, teaches how to build a language model end to end: tokenization, PyTorch and resource accounting, architectures, GPUs and kernels, parallelism, scaling laws, inference, evaluation, data, and alignment with RL.
It's one of the most hands-on LLM courses available, with demanding assignments that implement each piece.
Who it's for: Strong programmers with deep learning experience who want to understand and build LLMs from the ground up.
How long does it take?
The full playlist is 22.3 hours across 17 videos. Here's roughly how many days that is at different daily watch times and speeds, counting video time only:
| Daily time | At 1x | At 1.5x | At 2x |
|---|---|---|---|
| 30 min a day | 45 days | 30 days | 23 days |
| 1 h a day | 23 days | 15 days | 12 days |
| 2 h a day | 12 days | 8 days | 6 days |
Make your own schedule
The playlist link is already filled in. Press Plan, set how much time you have each day and your speed, and you'll get a day-by-day schedule with a finish date and a calendar file. It reads the live playlist, so new videos are included.
Week-by-week schedule
A sample plan at 1 hour a day and normal speed, using whole videos so none is split across days. That's 3 weeks in total. For a different pace, start date or speed, use the planner above.
Week 1Videos 1–7 · 9h 27m
Day 1 · 1h 19m
- 1Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 1: Overview and Tokenization1h 19m
Day 2 · 1h 19m
- 2Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lec. 2: Pytorch, Resource Accounting1h 19m
Day 3 · 1h 27m
- 3Stanford CS336 Lang. Modeling from Scratch | Spring 2025 | Lec. 3: Architectures, Hyperparameters1h 27m
Day 4 · 1h 22m
- 4Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 4: Mixture of experts1h 22m
Day 5 · 1h 14m
- 5Stanford CS336 I Language Modeling from Scratch | Spring 2025 | Lecture 5: GPUs1h 14m
Day 6 · 1h 20m
- 6Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 6: Kernels, Triton1h 20m
Day 7 · 1h 25m
- 7Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 7: Parallelism 11h 25m
Week 2Videos 8–14 · 9h 1m
Day 8 · 1h 15m
- 8Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 8: Parallelism 21h 15m
Day 9 · 1h 5m
- 9Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 9: Scaling laws 11h 5m
Day 10 · 1h 23m
- 10Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 10: Inference1h 23m
Day 11 · 1h 18m
- 11Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 11: Scaling laws 21h 18m
Day 12 · 1h 21m
- 12Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 12: Evaluation1h 21m
Day 13 · 1h 19m
- 13Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 13: Data 11h 19m
Day 14 · 1h 19m
- 14Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 14: Data 21h 19m
Week 3Videos 15–17 · 3h 52m
Day 15 · 1h 15m
- 15Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 15: Alignment - SFT/RLHF1h 15m
Day 16 · 1h 21m
- 16Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 16: Alignment - RL 11h 21m
Day 17 · 1h 16m
- 17Stanford CS336 Language Modeling from Scratch | Spring 2025 | Lecture 17: Alignment - RL 21h 16m
Tips for finishing
- The assignments are the course; budget several times the lecture time for them.
- Watch Karpathy's Zero to Hero first if you haven't built a transformer before.