Best AI Models for Linux - Page 5
View:
Compare the Top AI Models for Linux as of November 2026 - Page 5
Sort By:
-
1
StarCoder
BigCode
StarCoder and StarCoderBase are Large Language Models for Code (Code LLMs) trained on permissively licensed data from GitHub, including from 80+ programming languages, Git commits, GitHub issues, and Jupyter notebooks. Similar to LLaMA, we trained a ~15B parameter model for 1 trillion tokens. We fine-tuned StarCoderBase model for 35B Python tokens, resulting in a new model that we call StarCoder. We found that StarCoderBase outperforms existing open Code LLMs on popular programming benchmarks and matches or surpasses closed models such as code-cushman-001 from OpenAI (the original Codex model that powered early versions of GitHub Copilot). With a context length of over 8,000 tokens, the StarCoder models can process more input than any other open LLM, enabling a wide range of interesting applications. For example, by prompting the StarCoder models with a series of dialogues, we enabled them to act as a technical assistant.Starting Price: Free -
2
Llama 2
Meta
The next generation of our open source large language model. This release includes model weights and starting code for pretrained and fine-tuned Llama language models — ranging from 7B to 70B parameters. Llama 2 pretrained models are trained on 2 trillion tokens, and have double the context length than Llama 1. Its fine-tuned models have been trained on over 1 million human annotations. Llama 2 outperforms other open source language models on many external benchmarks, including reasoning, coding, proficiency, and knowledge tests. Llama 2 was pretrained on publicly available online data sources. The fine-tuned model, Llama-2-chat, leverages publicly available instruction datasets and over 1 million human annotations. We have a broad range of supporters around the world who believe in our open approach to today’s AI — companies that have given early feedback and are excited to build with Llama 2.Starting Price: Free -
3
Code Llama
Meta
Code Llama is a large language model (LLM) that can use text prompts to generate code. Code Llama is state-of-the-art for publicly available LLMs on code tasks, and has the potential to make workflows faster and more efficient for current developers and lower the barrier to entry for people who are learning to code. Code Llama has the potential to be used as a productivity and educational tool to help programmers write more robust, well-documented software. Code Llama is a state-of-the-art LLM capable of generating code, and natural language about code, from both code and natural language prompts. Code Llama is free for research and commercial use. Code Llama is built on top of Llama 2 and is available in three models: Code Llama, the foundational code model; Codel Llama - Python specialized for Python; and Code Llama - Instruct, which is fine-tuned for understanding natural language instructions.Starting Price: Free -
4
CogVideoX
CogVideoX
CogVideoX is a text-to-video generation tool. Before running the model, please refer to this guide to see how we use the GLM-4 model to optimize the prompt. This is crucial because the model is trained with long prompts, and a good prompt directly affects the quality of the generated video. Contains the inference code and fine-tuning code of SAT weights. It is recommended to improve based on the CogVideoX model structure. Innovative researchers use this code to better perform rapid stacking and development. A detailed wooden toy ship with intricately carved masts and sails is seen gliding smoothly over a plush, blue carpet that mimics the waves of the sea. The ship's hull is painted a rich brown, with tiny windows. The carpet, soft and textured, provides a perfect backdrop, resembling an oceanic expanse. Surrounding the ship are various other toys and children's items, hinting at a playful environment.Starting Price: Free -
5
TinyLlama
TinyLlama
The TinyLlama project aims to pretrain a 1.1B Llama model on 3 trillion tokens. With some proper optimization, we can achieve this within a span of "just" 90 days using 16 A100-40G GPUs. We adopted exactly the same architecture and tokenizer as Llama 2. This means TinyLlama can be plugged and played in many open-source projects built upon Llama. Besides, TinyLlama is compact with only 1.1B parameters. This compactness allows it to cater to a multitude of applications demanding a restricted computation and memory footprint.Starting Price: Free -
6
Pixtral Large
Mistral AI
Pixtral Large is a 124-billion-parameter open-weight multimodal model developed by Mistral AI, building upon their Mistral Large 2 architecture. It integrates a 123-billion-parameter multimodal decoder with a 1-billion-parameter vision encoder, enabling advanced understanding of documents, charts, and natural images while maintaining leading text comprehension capabilities. With a context window of 128,000 tokens, Pixtral Large can process at least 30 high-resolution images simultaneously. The model has demonstrated state-of-the-art performance on benchmarks such as MathVista, DocVQA, and VQAv2, surpassing models like GPT-4o and Gemini-1.5 Pro. Pixtral Large is available under the Mistral Research License for research and educational use, and under the Mistral Commercial License for commercial applications.Starting Price: Free -
7
Yi-Large
01.AI
Yi-Large is a proprietary large language model developed by 01.AI, offering a 32k context length with both input and output costs at $2 per million tokens. It stands out with its advanced capabilities in natural language processing, common-sense reasoning, and multilingual support, performing on par with leading models like GPT-4 and Claude3 in various benchmarks. Yi-Large is designed for tasks requiring complex inference, prediction, and language understanding, making it suitable for applications like knowledge search, data classification, and creating human-like chatbots. Its architecture is based on a decoder-only transformer with enhancements such as pre-normalization and Group Query Attention, and it has been trained on a vast, high-quality multilingual dataset. This model's versatility and cost-efficiency make it a strong contender in the AI market, particularly for enterprises aiming to deploy AI solutions globally.Starting Price: $0.19 per 1M input token -
8
BitNet
Microsoft
The BitNet b1.58 2B4T is a cutting-edge 1-bit Large Language Model (LLM) developed by Microsoft, designed to enhance computational efficiency while maintaining high performance. This model, built with approximately 2 billion parameters and trained on 4 trillion tokens, uses innovative quantization techniques to optimize memory usage, energy consumption, and latency. The platform supports multiple modalities and is particularly valuable for applications in AI-powered text generation, offering substantial efficiency gains compared to full-precision models.Starting Price: Free -
9
Gemma 3n
Google DeepMind
Gemma 3n is our state-of-the-art open multimodal model, engineered for on-device performance and efficiency. Made for responsive, low-footprint local inference, Gemma 3n empowers a new wave of intelligent, on-the-go applications. It analyzes and responds to combined images and text, with video and audio coming soon. Build intelligent, interactive features that put user privacy first and work reliably offline. Mobile-first architecture, with a significantly reduced memory footprint. Co-designed by Google's mobile hardware teams and industry leaders. 4B active memory footprint with the ability to create submodels for quality-latency tradeoffs. Gemma 3n is our first open model built on this groundbreaking, shared architecture, allowing developers to begin experimenting with this technology today in an early preview. -
10
GigaChat 3 Ultra
Sberbank
GigaChat 3 Ultra is a 702-billion-parameter Mixture-of-Experts model built from scratch to deliver frontier-level reasoning, multilingual capability, and deep Russian-language fluency. It activates just 36 billion parameters per token, enabling massive scale with practical inference speeds. The model was trained on a 14-trillion-token corpus combining natural, multilingual, and high-quality synthetic data to strengthen reasoning, math, coding, and linguistic performance. Unlike modified foreign checkpoints, GigaChat 3 Ultra is entirely original—giving developers full control, modern alignment, and a dataset free of inherited limitations. Its architecture leverages MoE, MTP, and MLA to match open-source ecosystems and integrate easily with popular inference and fine-tuning tools. With leading results on Russian benchmarks and competitive performance on global tasks, GigaChat 3 Ultra represents one of the largest and most capable open-source LLMs in the world.Starting Price: Free -
11
Composer 1.5
Cursor
Composer 1.5 is the latest agentic coding model from Cursor that balances speed and intelligence for everyday code tasks by scaling reinforcement learning approximately 20x more than its predecessor, enabling stronger performance on real-world programming challenges. It’s designed as a “thinking model” that generates internal reasoning tokens to analyze a user’s codebase and plan next steps, responding quickly to simple problems and engaging deeper reasoning on complex ones, while remaining interactive and fast for daily development workflows. To handle long-running tasks, Composer 1.5 introduces self-summarization, allowing the model to compress and carry forward context when it reaches context limits, which helps maintain accuracy across varying input lengths. Internal benchmarks show it surpasses Composer 1 in coding tasks, especially on more difficult issues, making it more capable for interactive use within Cursor’s environment. -
12
Qwen3.6
Alibaba
Qwen3.6 is a large language model developed by Alibaba as part of its Qwen AI model family, designed for real-world applications and advanced reasoning tasks. It focuses on improving stability, usability, and performance compared to earlier versions. The model supports multimodal capabilities, allowing it to process and reason across text, images, and other data types. Qwen3.6 is particularly strong in coding and developer workflows, offering improved accuracy for complex programming tasks. It uses a mixture-of-experts architecture, enabling efficient performance while maintaining large-scale model capabilities. The model is designed to be deployable in production environments, including enterprise and cloud-based systems. It can be integrated into applications or run locally using open-weight variants. Overall, Qwen3.6 delivers a powerful, efficient, and versatile AI solution for modern use cases.Starting Price: Free -
13
Lumen Outpost
Cosine
Lumen Outpost is Cosine’s targeted post-trained coding model, benchmarked against Kimi K2.6, its base model, GPT-5.5, GPT-5.4, and Gemini 3.1 Pro on highly complex, long-horizon coding tasks across 13 programming languages. The model is specialized not only for raw coding accuracy, but also for behavioral signals that matter in professional engineering workflows, including agent initiative, planning, scope discipline, action alignment, concise updates, and useful communication. Cosine’s benchmark report shows that highly targeted post-training transformed the base model’s capabilities, with Lumen Outpost outperforming Kimi K2.6 across Niche-Bench, Slop-Bench, Vibe-Bench, and cost per successful task. On Niche-Bench, an internal evaluation for niche, legacy, and environment-constrained programming languages, Lumen Outpost achieved a 53.9% score and led or tied in 9 of 13 assessed languages, with notable gains in Fortran, ABAP, Java, and Rust.Starting Price: $20 per month -
14
Cohere Parse
Cohere AI
Cohere Parse is a high-throughput vision-language model for processing large volumes of enterprise documents and converting complex, multimodal files into structured, machine-readable data. It goes beyond traditional OCR by understanding tables, forms, diagrams, embedded images, and document structure, returning clean Markdown for downstream processing and applications. Parse is trained for business documents across major industries and domains, including finance, insurance, and scientific work, and supports documents and images across nine major world languages. Spatial awareness preserves important visual relationships by returning bounding boxes for visual elements, helping improve retrieval, grounding, and automation. The model is designed for production-scale workloads, with high throughput and consistent parsing quality as document volumes grow. It can be used for automated document processing, extracting structured information from claims, contracts, and invoices. -
15
Nemotron 3.5 Transcribe
NVIDIA
NVIDIA Nemotron 3.5 Transcribe is an open-weight speech and audio model for developers building transcription and audio-processing systems. It converts recordings into text and supports verbatim or edited transcription, contextual vocabulary, speaker attribution, word-level timestamps, speech translation, language identification, and audio question answering. Its language matrix covers 40 languages, with task coverage varying by language. The model can process recordings up to one hour and is designed for offline inference using compatible NVIDIA GPU hardware and supported runtimes. Teams can integrate it into speech interfaces, meeting and media processing, customer-service analytics, translation tools, and other audio-understanding applications. Model weights and a fine-tuning toolkit are released under the OpenMDW-1.1 license; hardware and serving requirements depend on deployment. -
16
PaLM
Google
PaLM API is an easy and safe way to build on top of our best language models. Today, we’re making an efficient model available, in terms of size and capabilities, and we’ll add other sizes soon. The API also comes with an intuitive tool called MakerSuite, which lets you quickly prototype ideas and, over time, will have features for prompt engineering, synthetic data generation and custom-model tuning — all supported by robust safety tools. Select developers can access the PaLM API and MakerSuite in Private Preview today, and stay tuned for our waitlist soon. -
17
PaLM 2
Google
PaLM 2 is our next generation large language model that builds on Google’s legacy of breakthrough research in machine learning and responsible AI. It excels at advanced reasoning tasks, including code and math, classification and question answering, translation and multilingual proficiency, and natural language generation better than our previous state-of-the-art LLMs, including PaLM. It can accomplish these tasks because of the way it was built – bringing together compute-optimal scaling, an improved dataset mixture, and model architecture improvements. PaLM 2 is grounded in Google’s approach to building and deploying AI responsibly. It was evaluated rigorously for its potential harms and biases, capabilities and downstream uses in research and in-product applications. It’s being used in other state-of-the-art models, like Med-PaLM 2 and Sec-PaLM, and is powering generative AI features and tools at Google, like Bard and the PaLM API. -
18
Gemma
Google
Gemma is a family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models. Developed by Google DeepMind and other teams across Google, Gemma is inspired by Gemini, and the name reflects the Latin gemma, meaning “precious stone.” Accompanying our model weights, we’re also releasing tools to support developer innovation, foster collaboration, and guide the responsible use of Gemma models. Gemma models share technical and infrastructure components with Gemini, our largest and most capable AI model widely available today. This enables Gemma 2B and 7B to achieve best-in-class performance for their sizes compared to other open models. And Gemma models are capable of running directly on a developer laptop or desktop computer. Notably, Gemma surpasses significantly larger models on key benchmarks while adhering to our rigorous standards for safe and responsible outputs. -
19
Gemma 2
Google
A family of state-of-the-art, light-open models created from the same research and technology that were used to create Gemini models. These models incorporate comprehensive security measures and help ensure responsible and reliable AI solutions through selected data sets and rigorous adjustments. Gemma models achieve exceptional comparative results in their 2B, 7B, 9B, and 27B sizes, even outperforming some larger open models. With Keras 3.0, enjoy seamless compatibility with JAX, TensorFlow, and PyTorch, allowing you to effortlessly choose and change frameworks based on task. Redesigned to deliver outstanding performance and unmatched efficiency, Gemma 2 is optimized for incredibly fast inference on various hardware. The Gemma family of models offers different models that are optimized for specific use cases and adapt to your needs. Gemma models are large text-to-text lightweight language models with a decoder, trained in a huge set of text data, code, and mathematical content. -
20
FLUX1.1 Pro
Black Forest Labs
The FLUX1.1 Pro from Black Forest Labs sets a new benchmark in AI-powered image generation, delivering remarkable improvements in both speed and quality. This next-gen model outperforms its predecessor, FLUX.1 Pro, by being six times faster while enhancing image fidelity, prompt accuracy, and creative diversity. Key innovations include ultra-high-resolution rendering up to 4K and a Raw Mode for more natural, organic visuals. Available via the BFL API and integrated with platforms like Replicate and Freepik, FLUX1.1 Pro is the ultimate solution for professionals seeking advanced, scalable AI-generated imagery.Starting Price: Free -
21
HunyuanCustom
Tencent
HunyuanCustom is a multi-modal customized video generation framework that emphasizes subject consistency while supporting image, audio, video, and text conditions. Built upon HunyuanVideo, it introduces a text-image fusion module based on LLaVA for enhanced multi-modal understanding, along with an image ID enhancement module that leverages temporal concatenation to reinforce identity features across frames. To enable audio- and video-conditioned generation, it further proposes modality-specific condition injection mechanisms, an AudioNet module that achieves hierarchical alignment via spatial cross-attention, and a video-driven injection module that integrates latent-compressed conditional video through a patchify-based feature-alignment network. Extensive experiments on single- and multi-subject scenarios demonstrate that HunyuanCustom significantly outperforms state-of-the-art open and closed source methods in terms of ID consistency, realism, and text-video alignment. -
22
gpt-oss-20b
OpenAI
gpt-oss-20b is a 20-billion-parameter, text-only reasoning model released under the Apache 2.0 license and governed by OpenAI’s gpt-oss usage policy, built to enable seamless integration into custom AI workflows via the Responses API without reliance on proprietary infrastructure. Trained for robust instruction following, it supports adjustable reasoning effort, full chain-of-thought outputs, and native tool use (including web search and Python execution), producing structured, explainable answers. Developers must implement their own deployment safeguards, such as input filtering, output monitoring, and usage policies, to match the system-level protections of hosted offerings and mitigate risks from malicious or unintended behaviors. Its open-weight design makes it ideal for on-premises or edge deployments where control, customization, and transparency are paramount. -
23
gpt-oss-120b
OpenAI
gpt-oss-120b is a reasoning model engineered for deep, transparent thinking, delivering full chain-of-thought explanations, adjustable reasoning depth, and structured outputs, while natively invoking tools like web search and Python execution via the API. Built to slot seamlessly into self-hosted or edge deployments, it eliminates dependence on proprietary infrastructure. Although it includes default safety guardrails, its open-weight architecture allows fine-tuning that could override built-in controls, so implementers are responsible for adding input filtering, output monitoring, and governance measures to achieve enterprise-grade security. As a community–driven model card rather than a managed service spec, it emphasizes transparency, customization, and the need for downstream safety practices. -
24
MiMo-V2.5-Pro
Xiaomi Technology
Xiaomi MiMo-V2.5-Pro is an advanced open-source AI model designed to handle complex, long-horizon tasks with strong agentic capabilities. It features a Mixture-of-Experts architecture with over one trillion parameters and a large context window of up to one million tokens. The model is built to perform sophisticated reasoning, coding, and problem-solving across extended workflows. It demonstrates high performance on benchmark tests related to software engineering, reasoning, and general intelligence. MiMo-V2.5-Pro can autonomously complete complex projects, such as building full software systems or optimizing engineering designs. It uses hybrid attention mechanisms to balance efficiency and performance across long contexts. The model is also optimized for token efficiency, reducing computational cost while maintaining strong results. By combining scalability, efficiency, and advanced reasoning, MiMo-V2.5-Pro represents a major step forward in open-source AI models. -
25
MiMo-V2.5
Xiaomi Technology
Xiaomi MiMo-V2.5 is an advanced open-source AI model designed to combine strong agentic capabilities with native multimodal understanding. It can process and reason across text, images, and audio within a single unified system. The model uses a sparse Mixture-of-Experts architecture with hundreds of billions of parameters for efficient performance. It supports an extended context window of up to one million tokens, enabling long and complex workflows. MiMo-V2.5 is built to handle tasks such as coding, reasoning, and multimodal analysis with high accuracy. It incorporates dedicated visual and audio encoders to enhance perception and cross-modal reasoning. The model demonstrates strong benchmark performance across coding, reasoning, and multimodal tasks. By combining multimodality, efficiency, and agentic intelligence, MiMo-V2.5 advances the capabilities of open-source AI systems. -
26
North Mini Code
Cohere
North Mini Code is Cohere’s first agentic coding model for developers and the inaugural member of its next generation of powerful models. Small, efficient, and open-source, it is built for the sovereign developer ecosystem and designed to deliver strong software development performance without requiring extensive hardware. North Mini Code is a mixture-of-experts model with 30B total parameters and 3B active parameters, giving developers access to agentic coding capabilities in a compact and efficient form. The model is optimized for code generation, agentic software engineering, and terminal tasks, with a 256K total context length and up to 64K maximum generation. It is built for real-world developer workflows, including understanding and orchestrating sub-agents, mapping system architecture, running code reviews, and supporting coding agents that need to reason through complex software tasks. -
27
Big Pickle
OpenCode
Big Pickle is an AI model available through OpenCode Zen, a curated model provider focused on coding-agent workflows. The model is designed for text-based input, reasoning tasks, function calling, and developer workflows that require long-context understanding. Big Pickle supports a large context window, making it useful for working across bigger codebases, project files, technical prompts, and multi-step coding tasks. It can be accessed through OpenCode Zen using an OpenAI-compatible API format, allowing developers to integrate it into agentic coding tools and automation workflows. The model is positioned as a free or low-cost option within OpenCode’s coding-agent ecosystem. Big Pickle helps developers experiment with AI-assisted coding, reasoning, tool use, and long-context automation without relying only on premium frontier models.Starting Price: Free -
28
LongCat-2.0
LongCat
LongCat-2.0 is a 1.6 trillion total-parameter Mixture-of-Experts language model built on AI ASIC superpods, with about 48 billion parameters activated per token and strong performance across coding and agentic tasks. It is a substantial step up from previous LongCat models, combining large-scale sparse architecture with dedicated post-training for real-world software engineering, tool use, long-context reasoning, and multi-step agent workflows. LongCat-2.0 is trained and deployed entirely on AI ASIC superpods, with pretraining spanning more than 35 trillion tokens and millions of accelerator-hours, demonstrating frontier-scale training on alternative hardware platforms. To strengthen long-horizon tasks, the model introduces LongCat Sparse Attention and is trained on hundreds of billions of tokens of 1M-context data, giving it native support for ultra-long context tasks and reliable long-document understanding. -
29
Beam
Reflection
Beam is Reflection’s first open-weight model, a sparse Mixture-of-Experts model with 501 billion total parameters and 23 billion active parameters, built for coding, reasoning, and agentic workloads. Its capabilities come from large-scale pretraining and reinforcement learning, including training on 23.8 trillion diverse, curated, high-quality tokens from the web, public sources, and proprietary licensed datasets. Beam was trained with a particular focus on coding and agentic performance and is designed to deliver competitive open-weight capabilities with efficient inference compute. It supports complex software engineering, terminal, STEM, web search, tool-use, and general knowledge tasks, with reinforcement learning designed to improve multi-step reasoning, tool use, and adaptation to environment feedback. Users can control the tradeoff between performance and token usage through a reasoning effort parameter. -
30
ERNIE 3.0 Titan
Baidu
Pre-trained language models have achieved state-of-the-art results in various Natural Language Processing (NLP) tasks. GPT-3 has shown that scaling up pre-trained language models can further exploit their enormous potential. A unified framework named ERNIE 3.0 was recently proposed for pre-training large-scale knowledge enhanced models and trained a model with 10 billion parameters. ERNIE 3.0 outperformed the state-of-the-art models on various NLP tasks. In order to explore the performance of scaling up ERNIE 3.0, we train a hundred-billion-parameter model called ERNIE 3.0 Titan with up to 260 billion parameters on the PaddlePaddle platform. Furthermore, We design a self-supervised adversarial loss and a controllable language modeling loss to make ERNIE 3.0 Titan generate credible and controllable texts.
