AI Radar

Llama 3.1 Sonar

Retired · Language

Llama 3.1 Sonar covers Perplexity API language models launched in July 2024: llama-3.1-sonar-small-128k and llama-3.1-sonar-large-128k as online (search) and chat ids, joined by August 2024 by an online-only huge model listed at 405B parameters. The online ids were removed on 22 February 2025 in favour of Sonar and Sonar Pro. [1] [2] [3] Primary source

Timeline of Llama 3.1 Sonar →

Claims and evidence

  • Released The live changelog gives the month only. The archived post’s createdAt is 2024-07-31T04:05Z, which is 30 July in US time, so no day is recorded. Primary source[1]July 2024, Introducing New and Improved Sonar Models [2]
  • Deprecated Month of the notice for the online ids. No notice was found for the chat ids. Primary source[1]January 2025, API model deprecation notice
  • Retired Applies to the online ids (small, large, huge). The chat ids were dropped from the model cards between the captures of 2024-11-18 and 2024-12-21 (see the variants); no dated removal notice was found. Primary source[1]January 2025, API model deprecation notice
  • Status Retired Primary source[1]January 2025, API model deprecation notice
  • Successor of Llama 3 SonarThe launch notice says the new models surpass the previous iteration and retires the llama-3-sonar ids in favour of the Llama 3.1 family. Primary source[1]July 2024, Introducing New and Improved Sonar Models [2]
  • Change · Context length Context length grew to 131,072 tokens for the chat ids and 127,072 for the online ids, from 32768 and 28000.Compared with Llama 3 Sonar Archived primaryOriginal source unavailable[3]Context Length [4]Perplexity Models table, Context Length
  • Change · Size Adds a third, online-only model, llama-3.1-sonar-huge-128k, listed at 405B parameters; it was on the API docs by August 2024.Compared with Llama 3 Sonar Archived primaryOriginal source unavailable[3]Perplexity Sonar Models table
  • Change · Size The small model is listed at 8B parameters, where the Llama 3 Sonar small model was listed at 7B.Compared with Llama 3 Sonar Archived primaryOriginal source unavailable[3]Parameter Count [4]Perplexity Models table, Parameter Count
  • Change · Other Perplexity says the new models surpass the performance of the previous iteration.Compared with Llama 3 Sonar Archived primaryOriginal source unavailable[2]
  • Feature RetrievalThe Sonar (online) models have a search subsystem. Archived primaryOriginal source unavailable[3]note under Perplexity Sonar Models
  • Access API Archived primaryOriginal source unavailable[2]

Lineage

Predecessors

Successors

Based on

Not derived from another model.

Variants and derived

None recorded.

Siblings

None recorded.

All ancestors

All descendants

Variants

llama-3.1-sonar-small-128k

Same dates as Llama 3.1 Sonar.

  • Variant Inline variant in this record.The chat id is on the model cards captured on 2024-11-18 and gone from those captured on 2024-12-21. Primary source[2] [3] [5]Perplexity Chat Models [6]

Also known as: llama-3.1-sonar-small-128k-online, llama-3.1-sonar-small-128k-chat

Differs in:

  • Context window: 131,072 tokens Evidence not assessed [3]Context Length
  • Parameters: 8B Evidence not assessed [3]Parameter Count

llama-3.1-sonar-large-128k

Same dates as Llama 3.1 Sonar.

  • Variant Inline variant in this record.The chat id is on the model cards captured on 2024-11-18 and gone from those captured on 2024-12-21. Primary source[2] [3] [5]Perplexity Chat Models [6]

Also known as: llama-3.1-sonar-large-128k-online, llama-3.1-sonar-large-128k-chat

Differs in:

  • Context window: 131,072 tokens Evidence not assessed [3]Context Length
  • Parameters: 70B Evidence not assessed [3]Parameter Count

llama-3.1-sonar-huge-128k

  • Released Not in the July 2024 launch notice; first listed on the model cards captured on 2024-08-19, the only capture between 2024-07-25 and that date. No dated announcement was found. Archived primaryOriginal source unavailable[2] [3]Perplexity Sonar Models table
  • Retired Primary source[1]January 2025, API model deprecation notice
  • Variant Inline variant in this record.Online id only; no chat version was listed. Archived primaryOriginal source unavailable[3]

Also known as: llama-3.1-sonar-huge-128k-online

Differs in:

  • Context window: 127,072 tokens Evidence not assessed [3]Context Length
  • Parameters: 405B Evidence not assessed [3]Parameter Count

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Perplexity on AI Radar →