NLP Lineage

23 July 2024 · Scale and data

Llama 3.1 405B

The Llama 3 Herd of Models

Meta AI (Grattafiori, Dubey et al.)

Date
23 July 2024
What the date marks
Model release
Kind
Model
Organisation
Meta AI
Access
Llama 3.1 Community License
Other dates
paper 2024-07-31
Threads
In the guided tour
Chapter 8. The weights get out
Learning phases
6. Local models and deployment; 8. Training from scratch

A milestone: 2 later milestones descend directly from it, and its descendants reach 7 of the 10 threads. How levels are decided

Read the primary source (arxiv.org)

Full summary in progress. Until then, this page shows the milestone's record from the research data.

What it changed

Before it

Open weights trailed the closed frontier by a clear margin.

The problem

Nobody had released a model that could stand beside GPT-4 class systems.

What it introduced

405B parameters, 128K context, a candid 90-page account of the training, and a licence permitting distillation from outputs.

Why it matters

Open weights reached the frontier, and the distillation clause seeded a generation of smaller models.

What modern models inherit

The expectation that open weights compete at the frontier, now contested.

What came next

Llama 4's MoE in 2025, then Meta's pivot to closed models in 2026.

Before and after

The milestones it descends from, and the milestones that descend from it. Work developed alongside it is not counted. Weak links are left out unless you include them.

Before: where it came from

Llama 3.1 405B has 45 documented ancestors across 9 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 19 turning points among them.

Llama 3.1 405B has 45 documented ancestors across 9 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

Llama 3.1 405B has 46 recorded ancestors, including indirect ones, across 9 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 19 turning points among them.

Llama 3.1 405B has 46 recorded ancestors, including indirect ones, across 9 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2024Llama 3.1 405BThis milestone.
  2. 2024Llama 3Llama 3.1 405B built directly on thisChapter 8. The weights get out
  3. 2023Llama 2 Turning pointLlama 3 built directly on thisChapter 8. The weights get out
  4. 2023Position interpolationLlama 3.1 405B was influenced by thisChapter 10. Reading more than you remember
  5. 2023DPOLlama 3.1 405B built directly on thisChapter 6. Teaching models what people want
  6. 2023Grouped-query attention Turning pointLlama 3.1 405B built directly on thisChapter 9. Making it run anywhere
  7. 2023LLaMA Turning pointPosition interpolation built directly on thisChapter 8. The weights get out
  8. 2022Chinchilla Turning pointLlama 3.1 405B was influenced by thisChapter 5. Scale becomes a science
  9. 2022InstructGPT Turning pointDPO built directly on thisChapter 6. Teaching models what people want
  10. 2021GopherChinchilla built directly on thisChapter 5. Scale becomes a science
  11. 2021FLANInstructGPT was influenced by thisChapter 6. Teaching models what people want
  12. 2021RoPE Turning pointPosition interpolation built directly on thisChapter 10. Reading more than you remember
  13. 2020RLHF for summarisation Turning pointDPO built directly on thisChapter 6. Teaching models what people want
  14. 2020GPT-3 Turning pointInstructGPT built directly on thisChapter 5. Scale becomes a science
  15. 2020Scaling laws Turning pointChinchilla built directly on thisChapter 5. Scale becomes a science
  16. 2019Multi-query attentionGrouped-query attention built directly on thisChapter 9. Making it run anywhere
  17. 2019T5 Turning pointFLAN was influenced by thisChapter 4. Pretrain once, use everywhere
  18. 2019C4 Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  19. 2019RLHF for LMsRLHF for summarisation built directly on thisChapter 6. Teaching models what people want
  20. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  21. 2019GPT-2 Turning pointScaling laws was influenced by thisChapter 4. Pretrain once, use everywhere
  22. 2019WebTextGPT-2 was influenced by thisChapter 4. Pretrain once, use everywhere
  23. 2019Transformer-XLRoPE was influenced by thisChapter 4. Pretrain once, use everywhere
  24. 2018BERT Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  25. 2018SentencePieceT5 built directly on thisChapter 4. Pretrain once, use everywhere
  26. 2018GPT-1GPT-2 built directly on thisChapter 4. Pretrain once, use everywhere
  27. 2018ELMoBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  28. 2018ULMFiTBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  29. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  30. 2017Transformer Turning pointScaling laws built directly on this
    RoPE built directly on this
    Multi-query attention built directly on this    Chapter 3. Attention, then only attention
  31. 2017RLHF originRLHF for summarisation built directly on thisChapter 6. Teaching models what people want
  32. 2016Layer normTransformer built directly on this
    GPT-2 built directly on this    Chapter 3. Attention, then only attention
  33. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  34. 2015LM pretraining Turning pointGPT-1 was influenced by thisChapter 2. Words become vectors
  35. 2015BPE tokenizationLLaMA built directly on thisChapter 2. Words become vectors
  36. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  37. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  38. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  39. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  40. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  41. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  42. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  43. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  44. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  45. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  46. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2024Llama 3.1 405BThis milestone.
  2. 2024Llama 3Llama 3.1 405B built directly on thisChapter 8. The weights get out
  3. 2023Llama 2 Turning pointLlama 3 built directly on thisChapter 8. The weights get out
  4. 2023Position interpolationLlama 3.1 405B was influenced by thisChapter 10. Reading more than you remember
  5. 2023DPOLlama 3.1 405B built directly on thisChapter 6. Teaching models what people want
  6. 2023Grouped-query attention Turning pointLlama 3.1 405B built directly on thisChapter 9. Making it run anywhere
  7. 2023LLaMA Turning pointPosition interpolation built directly on thisChapter 8. The weights get out
  8. 2022Chinchilla Turning pointLlama 3.1 405B was influenced by thisChapter 5. Scale becomes a science
  9. 2022InstructGPT Turning pointDPO built directly on thisChapter 6. Teaching models what people want
  10. 2021GopherChinchilla built directly on thisChapter 5. Scale becomes a science
  11. 2021FLANInstructGPT was influenced by thisChapter 6. Teaching models what people want
  12. 2021RoPE Turning pointPosition interpolation built directly on thisChapter 10. Reading more than you remember
  13. 2020RLHF for summarisation Turning pointDPO built directly on thisChapter 6. Teaching models what people want
  14. 2020GPT-3 Turning pointInstructGPT built directly on thisChapter 5. Scale becomes a science
  15. 2020Scaling laws Turning pointChinchilla built directly on thisChapter 5. Scale becomes a science
  16. 2019Multi-query attentionGrouped-query attention built directly on thisChapter 9. Making it run anywhere
  17. 2019T5 Turning pointFLAN was influenced by thisChapter 4. Pretrain once, use everywhere
  18. 2019C4 Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  19. 2019RLHF for LMsRLHF for summarisation built directly on thisChapter 6. Teaching models what people want
  20. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  21. 2019GPT-2 Turning pointScaling laws was influenced by thisChapter 4. Pretrain once, use everywhere
  22. 2019WebTextGPT-2 was influenced by thisChapter 4. Pretrain once, use everywhere
  23. 2019Transformer-XLRoPE was influenced by thisChapter 4. Pretrain once, use everywhere
  24. 2018BERT Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  25. 2018SentencePieceT5 built directly on thisChapter 4. Pretrain once, use everywhere
  26. 2018GPT-1GPT-2 built directly on thisChapter 4. Pretrain once, use everywhere
  27. 2018ELMoGPT-1 was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  28. 2018ULMFiTGPT-1 was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  29. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  30. 2017Transformer Turning pointScaling laws built directly on this
    RoPE built directly on this
    Multi-query attention built directly on this    Chapter 3. Attention, then only attention
  31. 2017RLHF originRLHF for summarisation built directly on thisChapter 6. Teaching models what people want
  32. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  33. 2016Layer normTransformer built directly on this
    GPT-2 built directly on this    Chapter 3. Attention, then only attention
  34. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  35. 2015LM pretraining Turning pointGPT-1 was influenced by thisChapter 2. Words become vectors
  36. 2015BPE tokenizationLLaMA built directly on thisChapter 2. Words become vectors
  37. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  38. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  39. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  40. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  41. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  42. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  43. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  44. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  45. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  46. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  47. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

Llama 3.1 405B has 9 documented descendants across 3 chapters, reaching forward to Qwen3.8 in 2026. Showing its direct descendants.

Llama 3.1 405B has 9 documented descendants across 3 chapters, reaching forward to Qwen3.8 in 2026. Showing all of them.

Llama 3.1 405B has 14 recorded descendants, including indirect ones, across 3 chapters, reaching forward to Qwen3.8 in 2026. Showing its direct descendants.

Llama 3.1 405B has 14 recorded descendants, including indirect ones, across 3 chapters, reaching forward to Qwen3.8 in 2026. Showing all of them.

  1. 2024Llama 3.1 405BThis milestone.
  2. 2024DeepSeek-V3was influenced by Llama 3.1 405BChapter 8. The weights get out
  3. 2025DeepSeek-R1built directly on DeepSeek-V3Chapter 12. Thinking longer
  4. 2025Llama 4succeeded Llama 3.1 405BChapter 8. The weights get out
  5. 2025Qwen3was influenced by DeepSeek-R1Chapter 8. The weights get out
  6. 2025DeepSeek-V3.2built directly on DeepSeek-V3Chapter 10. Reading more than you remember
  7. 2026Qwen3.5built directly on Qwen3Chapter 8. The weights get out
  8. 2026Muse Sparksucceeded Llama 4Chapter 8. The weights get out
  9. 2026DeepSeek-V4was influenced by DeepSeek-V3Chapter 8. The weights get out
  10. 2026Qwen3.8built directly on Qwen3.5Chapter 8. The weights get out
  1. 2024Llama 3.1 405BThis milestone.
  2. 2024DeepSeek-V3was influenced by Llama 3.1 405BChapter 8. The weights get out
  3. 2025DeepSeek-R1built directly on DeepSeek-V3Chapter 12. Thinking longer
  4. 2025Llama 4succeeded Llama 3.1 405BChapter 8. The weights get out
  5. 2025Qwen3was influenced by DeepSeek-R1Chapter 8. The weights get out
  6. 2025Kimi K2was influenced by DeepSeek-V3 (indirect influence)Chapter 8. The weights get out
  7. 2025gpt-osswas influenced by DeepSeek-R1 (indirect influence)Chapter 8. The weights get out
  8. 2025GLM-4.5was influenced by DeepSeek-R1 (indirect influence)Chapter 8. The weights get out
  9. 2025DeepSeek-V3.2built directly on DeepSeek-V3Chapter 10. Reading more than you remember
  10. 2026Qwen3.5built directly on Qwen3Chapter 8. The weights get out
  11. 2026Muse Sparksucceeded Llama 4Chapter 8. The weights get out
  12. 2026DeepSeek-V4was influenced by DeepSeek-V3Chapter 8. The weights get out
  13. 2026GLM-5.2built directly on GLM-4.5Chapter 8. The weights get out
  14. 2026Kimi K3built directly on Kimi K2Chapter 8. The weights get out
  15. 2026Qwen3.8built directly on Qwen3.5Chapter 8. The weights get out

Direct links and their sources

Where it came from

Built directly on

Influenced by

What it led to

Influenced

Replaced by

Terms