Deci AI
4 models across 2 families · Latest: DeciLM 7B (2024-01)
Automating neural architecture design
Deci AI's portfolio covers 4 active models across 2 current families, spanning general LLM work. Open a model detail page to compare provider routes and sourced benchmarks.
Covers 0 workload areas across 4 active tracked models; last verified 2026-05-19.
Use it for
- Teams evaluating general LLM work across this lab's releases
- Comparing model families before committing to a flagship
- Migration and pricing follow-ups across 4 tracked models
Do not use it for
- Choosing a hosting provider without opening a model page for price ladders
Active models
4
Current models from this lab, excluding deprecated ones
Active families
2
Current model families from this lab
Open catalog
4 open
4 open source / 0 open weights
Lowest output price
$0.070 /1M
Cheapest tracked output across active models, per 1M tokens
Latest dated release
2024-01-16
DeciLM 7B
Freshness
2026-05-19
Researched 60d ago
Information
Release cadence
Showing 4 recent dated releases (full timeline below). Latest: DeciLM 7B (2024-01-16).
Where this lab wins
Not enough capability or benchmark coverage yet to call strengths for this lab.
Flagship quality / price signal
Flagship: DeciLM 7B (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.
Deci AI is an Israeli AI research organization founded in 2019. Automating neural architecture design. Deci AI ships 2 model families totaling 4 models, with the most recent release DeciLM 7B in 2024-01. Notable families include DeciLM and DeciCoder. Use it as a stable reference for lab background, release coverage, and follow-up model pages as they are added. Researchers and evaluators can. View official API endpoints, benchmark performance, and coding/agent fit for every Deci AI model.
About
Deci AI, founded in Tel Aviv in 2019, is a leading force in the realm of AI and deep learning acceleration. The company focuses on developing innovative solutions that empower developers by enhancing, optimizing, and expediting the deployment of AI models across various platforms, including cloud, edge, and mobile devices. A hallmark of Deci AI's approach is its emphasis on efficiency and cost-effectiveness, aiming to alleviate the computational and financial demands typical of AI model development and deployment 15. At the heart of Deci AI's technological advancements is their Automated Neural Architecture Construction (AutoNAC) engine. This cutting-edge NAS-based technology automatically configures and refines deep learning model architectures, focusing on achieving an optimal blend of accuracy and processing speed. By customizing models according to the unique demands of specific data sets, tasks, and performance goals, AutoNAC advances computer vision and natural language processing (NLP) technologies, evident in models like YOLO-NAS, DeciBERT, and DeciSeg 6. Deci AI's impact on generative AI and large language models (LLMs) is significant. The company has crafted a plethora of proprietary, fine-tunable LLMs and introduced Infery LLM, a novel inference engine complemented with an AI inference cluster management solution. Their flagship model, Deci-Nano, showcases leading-edge language and reasoning proficiency, surpassing rival models from giants like Google and Mistral regarding speed and cost-efficiency. Infery LLM notably advances Deci-Nano's real-time latency while minimizing expenses, enhancing the overall efficiency of the model 7. Additionally, Deci AI's portfolio includes noteworthy generative AI models such as the open-source DeciLM-7B, DeciCoder in various scales, and DeciDiffusion 2.0. These models focus on efficiency and are readily accessible via Deci's platform. The company ensures robust data security and control for enterprises, offering deployment options through a Virtual Private Cloud (VPC), directly within data centers, or via their API 3. Deci AI's remarkable journey includes collaborations with industry titans like Intel, AWS, NVIDIA, and HP, earning recognition from institutions like Gartner and MLPerf. In a notable development, Deci AI was acquired by NVIDIA in April 2024 and now operates as an NVIDIA subsidiary. This acquisition emphasizes Deci's substantial influence on the AI sector and its potential to further bolster NVIDIA's AI innovations. The company's commitment to open-source development, evidenced by models like DeciCoder, further illustrates its dedication to advancing innovation and accessibility within the AI community 5.
Featured models
| Model | Released | Context | Input price ($/1M) | Output price ($/1M) | License | Openness |
|---|---|---|---|---|---|---|
| DeciLM 7B | 2024-01-16 | 8k | $0.52 | $0.67 | Apache 2.0 | Open source |
| DeciLM 6B | 2024-01-16 | 4k | - | - | Apache 2.0 | Open source |
| DeciCoder 6B | 2023-08-15 | 4k | - | - | Apache 2.0 | Open source |
Model families
Recent releases
- DeciLM 7B- 2024-01-16
- DeciLM 6B- 2024-01-16
- DeciCoder 6B- 2023-08-15
- DeciCoder 1B- 2023-08-15
FAQ
Who founded Deci AI and when?
Deci AI was founded in 2019 and is associated with Tel Aviv, Israel. NVIDIA (April 2024).
What models has Deci AI released?
Deci AI ships 4 models across 2 families: DeciLM and DeciCoder.
Is Deci AI's technology open source?
All tracked models are released under Open Source.
Where is Deci AI headquartered?
Deci AI is headquartered in Tel Aviv, Israel.
What is Deci AI known for?
Automating neural architecture design. Its most prominent tracked family is DeciLM.
How can I access Deci AI's models?
Deci AI's models are available via Microsoft Foundry.
Explore related pages
Last reviewed: 2026-05-19. Data sourced from public lab announcements and provider documentation.