LLM Reference
LMSYS Org

LMSYS Org

8 models across 1 family · Latest: Vicuna 13B 16K (2023-10)

Crowdsourced AI model benchmarking

ClassificationJSON / Tool useNon-Profit

LMSYS Org's portfolio covers 8 active models across 1 current family, spanning classification and json / tool use. Open a model detail page to compare provider routes and sourced benchmarks.

Covers 2 workload areas across 8 active tracked models; last verified 2026-04-19.

Use it for

  • Teams evaluating classification and json / tool use across this lab's releases
  • Comparing model families before committing to a flagship
  • Migration and pricing follow-ups across 8 tracked models

Do not use it for

  • Choosing a hosting provider without opening a model page for price ladders

Active models

8

Current models from this lab, excluding deprecated ones

Active families

1

Current model families from this lab

Open catalog

8 open

0 open source / 8 open weights

Lowest output price

$0.200 /1M

Cheapest tracked output across active models, per 1M tokens

Latest dated release

2023-10-23

Vicuna 13B 16K

Freshness

2026-04-19

Researched 90d ago

stale

Information

Founded2023
Berkeley, California, United States

Release cadence

Showing 5 recent dated releases (full timeline below). Latest: Vicuna 13B 16K (2023-10-23).

Where this lab wins

  • Classification: 6 tracked models with MMLU-class moderation/safety coverage.
  • JSON/tool-use: 6 tracked models with BFCL / Nexus strict-JSON routing coverage.

Flagship quality / price signal

Flagship: Vicuna 13B 16K (best sourced coding quality-per-dollar in this portfolio).

Quality-per-dollar unavailable for this flagship — benchmark coverage or output token pricing is still missing.

LMSYS Org is an American AI research organization founded in 2023. Crowdsourced AI model benchmarking. LMSYS Org ships 1 model family totaling 8 models, with the most recent release Vicuna 13B 16K in 2023-10. Notable families include Vicuna. Use it as a stable reference for lab background, release coverage, and follow-up model pages as they are added. Researchers and evaluators can scan. View official API endpoints, benchmark performance, and coding/agent fit for every LMSYS Org model.

About

LMSYS Org, an open research initiative based in Berkeley, California, is pioneering the democratization of large model technologies. Established through a collaboration among students and faculty from prestigious institutions such as UC Berkeley, Stanford, UCSD, and CMU, the organization focuses on making advanced tools in artificial intelligence accessible and scalable to all. Their efforts span both machine learning and systems research, emphasizing the training of large language models (LLMs) and the advancement of distributed systems to optimize training and inference processes. A hallmark of LMSYS Org is its commitment to open-source principles, as reflected in the availability of their code and model weights for public use. This dedication underpins many of their groundbreaking projects, like the Chatbot Arena. This innovative platform allows for live, community-driven evaluation of language models, inviting users worldwide to engage in a gamified experience of model comparison. This approach not only provides developers and researchers with valuable real-world feedback but also challenges the conventional static benchmarks by continuously offering a variety of user prompts for assessment. Moreover, LMSYS Org has made strides with its distinct offerings such as Vicuna, an open-source chatbot which rivals some commercial counterparts like ChatGPT in performance. Trained by fine-tuning the LLaMA model on user-generated conversations, Vicuna highlights the organization’s potential to drive significant progress in conversational agents. Additionally, the development of tools such as SGLang, a high-speed serving engine for LLMs and VLMs, and evaluation instruments like MT-Bench and Arena Hard Auto, reveal LMSYS Org’s broad impact in AI model enhancement and evaluation. The influence of LMSYS Org extends beyond just technical advancements. Their community-centric and transparent approach to AI research is pushing the industry towards more inclusive innovation. By opening doors to collaboration and broader access to powerful AI tools, LMSYS Org is not only refining the capabilities of generative AI and LLMs but also setting a new benchmark for how research in this field is conducted and shared globally. Their initiatives are paving the way for the development of more robust and beneficial AI technologies, reflecting their foundational mission to democratize AI advancements.

Featured models

ModelReleasedContextInput price ($/1M)Output price ($/1M)LicenseOpenness
Vicuna 13B 16K2023-10-2316k--Llama 2 CommunityOpen weights
Vicuna 13B2023-10-232k$0.10$0.50Llama 2 CommunityOpen weights
Vicuna 7B 16K2023-10-2316k--Llama 2 CommunityOpen weights

Model families

Recent releases

  1. Vicuna 13B 16K- 2023-10-23
  2. Vicuna 13B- 2023-10-23
  3. Vicuna 7B 16K- 2023-10-23
  4. Vicuna 7B- 2023-10-23
  5. Vicuna 13B V1.5 16K- 2023-10-23

FAQ

Who founded LMSYS Org and when?

LMSYS Org was founded in 2023 and is associated with Berkeley, California, United States.

What models has LMSYS Org released?

LMSYS Org ships 8 models across 1 family: Vicuna.

Is LMSYS Org's technology open source?

All tracked models are released under Open Weights.

Where is LMSYS Org headquartered?

LMSYS Org is headquartered in Berkeley, California, United States.

What is LMSYS Org known for?

Crowdsourced AI model benchmarking. Its most prominent tracked family is Vicuna.

How can I access LMSYS Org's models?

LMSYS Org's models are available via GCP Vertex AI, Replicate API, and Together AI.

Explore related pages

Last reviewed: 2026-04-19. Data sourced from public lab announcements and provider documentation.