Top DeepSWE Alternatives

GPT-5.6 Sol is a next-generation OpenAI model previewed as the flagship option in the GPT-5.6 family. The series includes Sol for the strongest capability, Terra for balanced everyday work, and Luna for faster, lower-cost use cases. GPT-5.6 Sol is built for demanding work across coding, agentic automation, biology, cybersecurity, research, and enterprise knowledge workflows. The model introduces a new max reasoning effort that allows it to spend more time reasoning through difficult problems. It also adds ultra mode, which coordinates subagents to help accelerate complex tasks that benefit from parallel or multi-agent execution. In coding workflows, GPT-5.6 Sol is designed for command-line tasks that require planning, iteration, testing, tool coordination, and long-horizon software engineering judgment. In biology workflows, it is positioned for genomics and quantitative-biology analysis where efficient reasoning over complex scientific tasks matters. In cybersecurity, GPT-5.6 Sol supports legitimate defensive work such as vulnerability discovery, patch development, debugging, security education, code review, and authorized testing. OpenAI describes GPT-5.6 Sol as more capable at helping users find and fix vulnerabilities than reliably carrying out end-to-end attacks under tested conditions. The model’s release is paired with a layered safeguard system that includes model-level refusals, real-time misuse classifiers, paused generation for higher-risk cases, account-level review, automated red-teaming, third-party testing, differentiated access, and enterprise safety controls. GPT-5.6 Sol helps developers, researchers, enterprises, and cyber defenders use frontier AI for advanced technical work while supporting safer deployment, stronger oversight, and phased access.

Gemini 4 Carbon is an advanced multimodal artificial intelligence model developed by Google DeepMind for cybersecurity research, software engineering, long-horizon autonomous agents, and complex analytical workflows. The model builds on Gemini 4 Argon with improvements in reasoning, coding, scientific tasks, multimodal understanding, and sustained execution of multistep engineering projects. Its cybersecurity capabilities cover the vulnerability management lifecycle, including identifying software weaknesses, verifying exploitability, and generating patches that address underlying security flaws. In Google's published evaluations, Gemini 4 Carbon achieved 88.2% on SEC-bench Pro for vulnerability discovery, 54.8% on ExploitGym for exploit validation, and 71.0% on CWE-bench for vulnerability remediation. Internal evaluations also showed more than twice the secure code generation success rate of Gemini 4 Argon. The model supports long-running agentic workflows that can investigate difficult bugs, analyze large software repositories, and perform iterative problem-solving across extended execution periods. Its software development capabilities include code generation, debugging, low-level systems analysis, full-stack web development, and automated implementation of complex engineering tasks. Gemini 4 Carbon incorporates a multilayered safety architecture with misuse prevention, prompt injection defenses, behavioral monitoring, and controls intended to reduce unauthorized actions. Google reports improvements in factual reliability, task completion, and instruction alignment, including approximately half as many unsupported claims compared with Gemini 4 Argon in internal evaluations.

MiMo-V2.6-Pro is Xiaomi MiMo’s flagship open-source omnimodal model for software engineering, agentic automation, multimodal reasoning, visual design, research, and creative production. The model was developed through large-scale reinforcement learning on heterogeneous tasks spanning coding, general agents, visual workflows, and cybersecurity. Xiaomi trained MiMo-V2.6-Pro across roughly 750,000 trajectories using large asynchronous batches, long-context training, multi-task environments, and expanded grader compute. The resulting model is designed to plan, execute, verify, and refine complex work across multiple tools and interaction environments. In software development, MiMo-V2.6-Pro supports long-horizon coding, terminal work, automation, debugging, and other agent-driven engineering tasks. Its multimodal capabilities allow it to generate interactive 3D worlds, create Blender assets from text or reference images, and control simulated robotic systems using continuous visual feedback. The model can also build frontend interfaces, design slide decks, work with Figma and media-generation tools, and automate portions of video production from concept through editing and narration. Creative capabilities extend to music composition, including generating arrangements, musical scores, and MIDI output. For research, MiMo-V2.6-Pro has been demonstrated performing literature searches, generating scientific hypotheses, running computational tools, screening materials, and assisting with formal mathematical proofs. Xiaomi has open-sourced the model family together with its technical report, reinforcement learning environments, and training code to support reproducibility and further research. MiMo-V2.6-Pro is available through Xiaomi MiMo’s desktop and developer products, OpenRouter, Hugging Face, and an API, with an UltraSpeed version offered for workflows that require much faster generation.

NVIDIA Nemotron 3.5 Super VL is an open multimodal artificial intelligence model designed to support specialized autonomous agents, enterprise automation, and complex reasoning tasks. Introduced in October 2026, the model combines language processing, visual understanding, software engineering capabilities, and external tool interactions within a unified architecture. It uses a 120-billion-parameter mixture-of-experts design with approximately 12 billion active parameters per token to optimize inference efficiency. The model was distilled from NVIDIA Nemotron 3 Ultra and incorporates additional training focused on coding, multimodal reasoning, and agentic task execution. Its vision-language capabilities support analysis of images, charts, tables, diagrams, multipage documents, and video content. Document intelligence workflows can extract structured information, interpret visual evidence, connect information across pages, and generate summaries or reports. Software engineering features enable the model to analyze repositories, identify programming errors, test code, and recommend modifications across development workflows. Its agentic capabilities support web research, external tool usage, multistep problem-solving, cybersecurity vulnerability analysis, and enterprise process automation. NVIDIA NeMo provides customization tools for supervised fine-tuning, low-rank adaptation, and reinforcement learning using proprietary datasets, organizational policies, and specialized workflows. The model incorporates LiLiCorr speculative decoding, multi-token prediction, context reuse, and NVFP4 quantization to improve inference performance on compatible hardware. Nemotron 3.5 Super VL is distributed with open weights and supports local, cloud, and data center deployment through NVIDIA infrastructure and tools such as NVIDIA NIM.

SWE-2 is Cognition’s coding model for software engineering agents, developed to improve the balance between capability, reasoning cost, and execution efficiency. The model is post-trained from Kimi K3, a multi-trillion-parameter model that had already received extensive reinforcement learning for agentic coding. Cognition further trained SWE-2 with a reinforcement learning algorithm that optimizes several reasoning-effort levels during a single training run. These effort levels let users trade off speed and cost against deeper planning, codebase exploration, and verification for more difficult assignments. SWE-2 is designed to reduce the over-exploration seen in earlier models by identifying relevant files and implementation paths more quickly. Its software engineering abilities include repository analysis, code writing and editing, debugging, testing, build and lint workflows, terminal tasks, and verification of completed work. The model places additional emphasis on writing end-to-end tests, catching edge cases and regressions, and gathering evidence instead of simply accepting assumptions in a prompt. Cognition’s training approach also uses cost penalties tied to the model’s performance frontier, length-weighted reward baselines, speculative decoding improvements, low-precision inference techniques, and expanded reinforcement learning data. Training data includes more diverse repositories, additional instruction-following requirements, and iterative verifier improvements designed to reduce reward hacking and false validation. SWE-2 is benchmarked against models such as GPT-6 Astra, GPT-5.6 Sol, Fable 5.1, Grok 4.6, and Kimi K3, with Cognition positioning it around strong coding performance at substantially lower cost. SWE-2 is intended for use across Cognition’s Devin ecosystem, including Desktop and CLI, with rollout to Devin Web and Fusion.

Qwen3.8-Max is a large-scale AI model from Qwen built for coding, coworking, research, long-horizon planning, and multimodal agent workflows. It is positioned as the most capable model in the Qwen family to date, with open weights announced for release after launch. The model uses a 2.4 trillion-parameter architecture with 95 billion active parameters and is available through QwenCloud. Qwen3.8-Max is designed to complete complex, open-ended goals end to end rather than only answer isolated prompts. In coding workflows, it can write and run code, create self-evolving harnesses, normalize requirements into issues, execute tasks through agents, run tests, trigger CI checks, and iterate through feedback. Its autonomous coding examples include a 10+ day project run, a research-paper reproduction and improvement loop, and a 24-hour online competition solution that beat most participating human teams. For professional work, Qwen3.8-Max is built to handle multi-step, tool-heavy workflows across compliance, design, food operations, engineering, rehabilitation, sports analytics, and quantitative research. The model also supports long-horizon decision-making, including autonomous chip-design optimization and extended e-commerce operations simulations. Its multimodal capabilities cover images, complex PDFs, long videos, visual production, interface inspection, frontend reconstruction, Blender visualization, interactive applications, and visual feedback loops. Qwen3.8-Max can be integrated through QwenCloud APIs and used with agent frameworks or coding assistants such as Claude Code, Codex, Qoder CLI, Qwen Code, and OpenClaw. By combining agentic coding, reasoning controls, multimodal understanding, visual self-correction, long-context workflows, tool use, and open-weight availability, Qwen3.8-Max helps developers and organizations build autonomous AI systems that can produce dependable deliverables.

Devstral 2 is an innovative, open-source AI model tailored for software engineering, transcending simple code suggestions to fully understand and manipulate entire codebases; this advanced functionality enables it to execute tasks such as multi-file edits, bug fixes, refactoring, managing dependencies, and generating code that is aware of its context. The suite includes a powerful 123-billion-parameter model alongside a streamlined 24-billion-parameter variant called “Devstral Small 2,” offering flexibility for teams; the larger model excels in handling intricate coding tasks that necessitate a deep contextual understanding, whereas the smaller model is optimized for use on less robust hardware. With a remarkable context window capable of processing up to 256 K tokens, Devstral 2 is adept at analyzing extensive repositories, tracking project histories, and maintaining a comprehensive understanding of large files, which is especially advantageous for addressing the challenges of real-world software projects. Additionally, the command-line interface (CLI) further enhances the model’s functionality by monitoring project metadata, Git statuses, and directory structures, thereby enriching the AI’s context and making “vibe-coding” even more impactful. This powerful blend of features solidifies Devstral 2's role as a revolutionary tool within the software development ecosystem, offering unprecedented support for engineers. As the landscape of software engineering continues to evolve, tools like Devstral 2 promise to redefine the way developers approach coding tasks.

Muse Spark 1.2 is a coding-focused AI model from Meta designed to support advanced software engineering tasks through Muse Code and the Meta Model API. The model builds on Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows. Muse Spark 1.2 powers Muse Code, a terminal coding agent that can plan repository changes, write code, validate outputs, and work across large codebases. Muse Code uses persistent async background agents that stay active throughout a session to reduce redundant information gathering and support difficult multi-step work. The runtime uses a local event log where model calls, tool runs, approvals, and edits are appended, making sessions replay-exact and restart-safe. Muse Spark 1.2 was co-trained with Muse Code so the model can take advantage of its toolset, harness workflows, goals, compaction, and subagent architecture. Meta significantly scaled training compute on coding tasks and expanded training environment diversity to improve the model’s engineering capabilities. The model was also trained on long-horizon coding tasks, including whole-repository generation, large end-to-end projects, auto-research, and extended iterative work. Its training approach uses planning, goal conditioning, context compaction, rejection-sampled harness trajectories, and self-improvement data generated with Muse Spark 1.1. Meta also tested Muse Spark 1.2 on long-running GPU kernel optimization workflows where the model wrote, compiled, profiled, and improved Triton kernels over many tool calls. By combining coding-focused training, agentic runtime integration, persistent subagents, long-horizon reasoning, replay-safe execution, and API availability, Muse Spark 1.2 helps developers and AI agents complete complex software engineering work with less intervention.

DeepCoder, a fully open-source initiative for code reasoning and generation, has been created through a collaboration between the Agentica Project and Together AI. Built on the foundation of DeepSeek-R1-Distilled-Qwen-14B, it has been fine-tuned using distributed reinforcement learning techniques, achieving an impressive accuracy of 60.6% on LiveCodeBench, which represents an 8% improvement compared to its predecessor. This remarkable performance positions it competitively alongside proprietary models such as o3-mini (2025-01-031 Low) and o1, all while operating with a streamlined 14 billion parameters. The training process was intensive, lasting 2.5 weeks on a fleet of 32 H100 GPUs and utilizing a meticulously curated dataset comprising around 24,000 coding challenges obtained from reliable sources such as TACO-Verified, PrimeIntellect SYNTHETIC-1, and submissions to LiveCodeBench. Each coding challenge was required to include a valid solution paired with at least five unit tests to ensure robustness during the reinforcement learning phase. Additionally, DeepCoder employs innovative methods like iterative context lengthening and overlong filtering to effectively handle long-range contextual dependencies, allowing it to tackle complex coding tasks with proficiency. This distinctive approach not only enhances DeepCoder's accuracy and reliability in code generation but also positions it as a significant player in the landscape of code generation models. As a result, developers can rely on its capabilities for diverse programming challenges.

Devstral Small 2 is a condensed, 24 billion-parameter variant of Mistral AI's groundbreaking coding-focused models, made available under the adaptable Apache 2.0 license to support both local use and API access. Alongside its more extensive sibling, Devstral 2, it offers "agentic coding" capabilities tailored for low-computational environments, featuring a substantial 256K-token context window that enables it to understand and alter entire codebases with ease. With a performance score nearing 68.0% on the widely recognized SWE-Bench Verified code-generation benchmark, Devstral Small 2 distinguishes itself within the realm of open-weight models that are much larger. Its compact structure and efficient design allow it to function effectively on a single GPU or even in CPU-only setups, making it an excellent option for developers, small teams, or hobbyists who may lack access to extensive data-center facilities. Moreover, despite being smaller, Devstral Small 2 retains critical functionalities found in its larger counterparts, such as the capability to reason through multiple files and adeptly manage dependencies, ensuring that users enjoy substantial coding support. This combination of efficiency and high performance positions it as an indispensable asset for the coding community. Additionally, its user-friendly approach ensures that both novice and experienced programmers can leverage its capabilities without significant barriers.

Qwen2.5-Max is a cutting-edge Mixture-of-Experts (MoE) model developed by the Qwen team, trained on a vast dataset of over 20 trillion tokens and improved through techniques such as Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF). It outperforms models like DeepSeek V3 in various evaluations, excelling in benchmarks such as Arena-Hard, LiveBench, LiveCodeBench, and GPQA-Diamond, and also achieving impressive results in tests like MMLU-Pro. Users can access this model via an API on Alibaba Cloud, which facilitates easy integration into various applications, and they can also engage with it directly on Qwen Chat for a more interactive experience. Furthermore, Qwen2.5-Max's advanced features and high performance mark a remarkable step forward in the evolution of AI technology. It not only enhances productivity but also opens new avenues for innovation in the field.

KAT-Coder is an advanced AI coding solution that goes beyond traditional autocomplete features by enabling a thorough software development workflow that incorporates reasoning, planning, and execution. This innovative system is recognized as the leading coding model in the KAT ecosystem, designed specifically for "agentic coding," which empowers the model to generate code snippets while also diagnosing issues, proposing solutions, performing tests, and refining various files throughout an ongoing development cycle. Through its seamless integration into developer environments via API endpoints and proxy layers compatible with tools like Claude Code, developers can retain their familiar workflows without the need to change their interfaces. KAT-Coder utilizes a sophisticated multi-stage training pipeline that merges supervised fine-tuning with extensive reinforcement learning, allowing it to understand programming contexts and effectively manage complex tasks. As a result, KAT-Coder significantly boosts productivity and equips developers with the freedom to concentrate on the more creative elements of their projects. Moreover, its adaptive capabilities ensure that developers can continuously improve their coding practices, which leads to even more innovative solutions.

Qwen3-Coder-Next is an open-weight language model designed specifically for coding agents and local development, excelling in complex coding reasoning, proficient tool utilization, and effectively managing long-term programming tasks with exceptional efficiency through a mixture-of-experts framework that balances strong capabilities with a resource-conscious design. This model significantly boosts the coding abilities of software developers, AI system designers, and automated coding systems, enabling them to create, troubleshoot, and understand code with a deep contextual insight while skillfully recovering from execution errors, making it particularly suitable for autonomous coding agents and development-focused applications. Additionally, Qwen3-Coder-Next offers remarkable performance comparable to models with larger parameters but operates with a reduced number of active parameters, making it a cost-effective solution for tackling complex and dynamic programming challenges in both research and production environments. Ultimately, this innovative model is designed to enhance the efficiency and effectiveness of the development process, paving the way for more agile and responsive software creation. Its ability to streamline workflows further underscores its potential to transform how programming tasks are approached and executed.

Qwen3-Coder is a multifaceted coding model available in different sizes, prominently showcasing the 480B-parameter Mixture-of-Experts variant with 35B active parameters, which adeptly manages 256K-token contexts that can be scaled up to 1 million tokens. It demonstrates remarkable performance comparable to Claude Sonnet 4, having been pre-trained on a staggering 7.5 trillion tokens, with 70% of that data comprising code, and it employs synthetic data fine-tuned through Qwen2.5-Coder to bolster both coding proficiency and overall effectiveness. Additionally, the model utilizes advanced post-training techniques that incorporate substantial, execution-guided reinforcement learning, enabling it to generate a wide array of test cases across 20,000 parallel environments, thus excelling in multi-turn software engineering tasks like SWE-Bench Verified without requiring test-time scaling. Beyond the model itself, the open-source Qwen Code CLI, inspired by Gemini Code, equips users to implement Qwen3-Coder within dynamic workflows by utilizing customized prompts and function calling protocols while ensuring seamless integration with Node.js, OpenAI SDKs, and environment variables. This robust ecosystem not only aids developers in enhancing their coding projects efficiently but also fosters innovation by providing tools that adapt to various programming needs. Ultimately, Qwen3-Coder stands out as a powerful resource for developers seeking to improve their software development processes.

DeepSeek-V4.1-Flash-1017 is an open-weight artificial intelligence model developed by DeepSeek for multimodal understanding, advanced reasoning, autonomous agent workflows, and software engineering. Introduced on October 17, 2026, the release updates DeepSeek-V4.1-Flash with improvements in reasoning accuracy, coding capabilities, agent performance, and computational efficiency. The model builds on DeepSeek's mixture-of-experts architecture and supports native visual understanding alongside natural language processing. Its one-million-token context window enables applications to process extensive documents, large code repositories, and long sequences of interactions. Developers can configure reasoning effort and use the model for tasks involving mathematical reasoning, programming, information analysis, and complex problem-solving. Software engineering capabilities include code generation, debugging, repository analysis, automated testing, and implementation of features across existing projects. DeepSeek also trained the model using reinforcement learning with self-feedback to improve its ability to create complete projects, including web applications, interactive games, and 3D environments. Its agentic capabilities support multi-turn conversations, external tool calling, task planning, and integration with automated development environments. DeepSeek reports improvements over the previous version on benchmarks including Terminal-Bench, DeepSWE, ProgramBench, Automation-Bench, and Agents' Last Exam. The model includes reference prompt encoding tools and supports deployment through Hugging Face Transformers, vLLM, SGLang, and compatible inference infrastructure. DeepSeek-V4.1-Flash-1017 is available through the DeepSeek API under the deepseek-flash model identifier and through downloadable model weights licensed under the MIT License.

Qwen3.6-27B stands as an open-source, dense multimodal language model within the Qwen3.6 lineup, crafted to deliver exceptional capabilities in coding, reasoning, and workflows driven by agents, all while utilizing a streamlined parameter count of 27 billion. This model is distinguished by its performance, often surpassing or closely rivaling larger models on critical benchmarks, especially in tasks that involve agent-based coding. It operates in two distinct modes—thinking and non-thinking—allowing it to adjust the depth of its reasoning and the speed of its responses to align with the specific demands of various tasks. Furthermore, it accommodates a broad range of input formats, which includes text, images, and video, demonstrating its adaptability. As an integral part of the Qwen3.6 series, this model emphasizes practical functionality, reliability, and the boost of developer efficiency, drawing on feedback from the community and the practical needs of real-world applications. Its forward-thinking design not only addresses current user requirements but also foresees future developments in the realm of artificial intelligence, ensuring that it remains relevant and effective over time. Thus, Qwen3.6-27B represents a significant step forward in the evolution of language models, integrating innovative features that enhance user interaction and streamline workflows.

Tülu 3 represents a state-of-the-art language model designed by the Allen Institute for AI (Ai2) with the objective of enhancing expertise in various domains such as knowledge, reasoning, mathematics, coding, and safety. Built on the foundation of the Llama 3 Base, it undergoes an intricate four-phase post-training process: meticulous prompt curation and synthesis, supervised fine-tuning across a diverse range of prompts and outputs, preference tuning with both off-policy and on-policy data, and a distinctive reinforcement learning approach that bolsters specific skills through quantifiable rewards. This open-source model is distinguished by its commitment to transparency, providing comprehensive access to its training data, coding resources, and evaluation metrics, thus helping to reduce the performance gap typically seen between open-source and proprietary fine-tuning methodologies. Performance evaluations indicate that Tülu 3 excels beyond similarly sized models, such as Llama 3.1-Instruct and Qwen2.5-Instruct, across multiple benchmarks, emphasizing its superior effectiveness. The ongoing evolution of Tülu 3 not only underscores a dedication to enhancing AI capabilities but also fosters an inclusive and transparent technological landscape. As such, it paves the way for future advancements in artificial intelligence that prioritize collaboration and accessibility for all users.

Qwen3.7-Max signifies the pinnacle of innovation in Qwen's proprietary model series, specifically designed for the agent-centric era, and acts as a solid platform for a multitude of applications such as writing and debugging code, automating office workflows, and sustaining prolonged autonomous browsing sessions. This model excels in coding performance, showcasing exceptional skills in software engineering, terminal operations, graphical user interface interactions, web surfing, and the effective use of agentic tools. By improving the synergy between the model's intelligence and actual agent execution, Qwen3.7-Max supports sophisticated planning, reasoning over extended contexts, reliable function invocation, and the management of complex, multi-step tasks in intricate workflows. Additionally, it enhances multimodal and document-oriented tasks via Qwen Studio, which facilitates chatbot interactions, interprets images and videos, creates visuals, processes documents, develops presentations, provides coding assistance, performs thorough research, and supports web development. With this extensive array of capabilities, Qwen3.7-Max is positioned as a premier solution for various operational requirements in today's dynamic digital environment, ensuring users can efficiently tackle a wide range of challenges. As technology continues to evolve, the importance of such advanced models will only grow, making Qwen3.7-Max an invaluable asset for future endeavors.

Qwen3, the latest large language model from the Qwen family, introduces a new level of flexibility and power for developers and researchers. With models ranging from the high-performance Qwen3-235B-A22B to the smaller Qwen3-4B, Qwen3 is engineered to excel across a variety of tasks, including coding, math, and natural language processing. The unique hybrid thinking modes allow users to switch between deep reasoning for complex tasks and fast, efficient responses for simpler ones. Additionally, Qwen3 supports 119 languages, making it ideal for global applications. The model has been trained on an unprecedented 36 trillion tokens and leverages cutting-edge reinforcement learning techniques to continually improve its capabilities. Available on multiple platforms, including Hugging Face and ModelScope, Qwen3 is an essential tool for those seeking advanced AI-powered solutions for their projects.

Qwen3.8-27B is an open-weights 27B-class model connected to Alibaba’s Qwen3.8 release, built for developers, researchers, and AI teams that need a capable but more deployable model size. Alibaba’s Qwen3.8 launch described the broader model family as optimized for coding and cowork scenarios, including software development, document processing, data analysis, and professional workflows. Reports state that Alibaba planned to open-source Qwen3.8-Max alongside Qwen3.8-27B, expanding access for developers and researchers. Qwen3.8-27B gives builders a smaller alternative to the 2.4T-parameter Qwen3.8-Max model, which third-party coverage describes as Qwen’s first Max-scale model planned for open weights. The model is well suited for coding assistance, local development, agent testing, workflow automation, data analysis, document understanding, and private AI experimentation. QwenCloud documentation lists Qwen3.8-Max as supporting a 1M context window, thinking, function calling, built-in tools, and structured output, showing the broader Qwen3.8 generation’s focus on advanced agent and application workflows. Qwen3.8-27B is especially useful for teams that want Qwen-family capabilities without the infrastructure demands of Max-scale deployment. Community posts around the release point to active interest in Hugging Face, Unsloth GGUF, Ollama, and local inference use cases. Third-party coverage also notes practical hardware discussions around quantized Qwen3.8-27B deployment, including claims that 4-bit variants can fit more easily on consumer or workstation GPUs. The model can be positioned for organizations that need open AI infrastructure, coding agents, local model evaluation, private deployments, and cost-controlled experimentation. By combining open-weight access, a practical 27B model size, Qwen3.8-era performance ambitions, coding-oriented workflows, and local deployment interest, Qwen3.8-27B gives developers a flexible foundation for building AI products and agents.

Qwen3.5 is an advanced open-weight multimodal AI system built to serve as the foundation for native digital agents capable of reasoning across text, images, and video. The primary release, Qwen3.5-397B-A17B, introduces a hybrid architecture that combines Gated DeltaNet linear attention with a sparse mixture-of-experts design, activating just 17 billion parameters per inference pass while maintaining a total parameter count of 397 billion. This selective activation dramatically improves decoding throughput and cost efficiency without sacrificing benchmark-level performance. Qwen3.5 demonstrates strong results across knowledge, multilingual reasoning, coding, STEM tasks, search agents, visual question answering, document understanding, and spatial intelligence benchmarks. The hosted Qwen3.5-Plus variant offers a default one-million-token context window and integrated tool usage such as web search and code interpretation for adaptive problem-solving. Expanded multilingual support now covers 201 languages and dialects, backed by a 250k vocabulary that enhances encoding and decoding efficiency across global use cases. The model is natively multimodal, using early fusion techniques and large-scale visual-text pretraining to outperform prior Qwen-VL systems in scientific reasoning and video analysis. Infrastructure innovations such as heterogeneous parallel training, FP8 precision pipelines, and disaggregated reinforcement learning frameworks enable near-text baseline throughput even with mixed multimodal inputs. Extensive reinforcement learning across diverse and generalized environments improves long-horizon planning, multi-turn interactions, and tool-augmented workflows. Designed for developers, researchers, and enterprises, Qwen3.5 supports scalable deployment through Alibaba Cloud Model Studio while paving the way toward persistent, economically aware, autonomous AI agents.

Laguna XS.2 stands out as Poolside's groundbreaking open-weight coding model, noted for being the lightest and fastest in the Laguna lineup. Equipped with a staggering 33 billion parameters organized in a Mixture of Experts structure, of which 3 billion are active, this model has undergone extensive training in-house utilizing 30 trillion tokens. As the most recent generation model available to the public, it features a second-generation architecture and represents Poolside's first open-weight release, benefiting from lessons learned during the Laguna M.1 training process, which utilized synthetic data and reinforcement learning. Tailored specifically to optimize agentic coding workflows, Laguna XS.2 is exceptional in coding, acting, and rapid iteration, particularly within Poolside's coding agent ecosystem. This model is especially beneficial for developers and teams in need of a lightweight and efficient coding solution, as opposed to more complex frontier systems. Released under the flexible Apache 2.0 license, it enables the community to evaluate, refine, quantize, and build upon its weights, fostering an environment of collaborative development. Ultimately, Laguna XS.2 not only serves as a powerful tool for agentic coding but also promotes creativity and experimentation among its users, allowing for a diverse range of applications and enhancements.

Sky-T1-32B-Preview represents a groundbreaking open-source reasoning model developed by the NovaSky team at UC Berkeley's Sky Computing Lab. It achieves performance levels similar to those of proprietary models like o1-preview across a range of reasoning and coding tests, all while being created for under $450, emphasizing its potential to provide advanced reasoning skills at a lower cost. Fine-tuned from Qwen2.5-32B-Instruct, this model was trained on a carefully selected dataset of 17,000 examples that cover diverse areas, including mathematics and programming. The training was efficiently completed in a mere 19 hours with the aid of eight H100 GPUs using DeepSpeed Zero-3 offloading technology. Notably, every aspect of this project—spanning data, code, and model weights—is fully open-source, enabling both the academic and open-source communities to not only replicate but also enhance the model's functionalities. Such openness promotes a spirit of collaboration and innovation within the artificial intelligence research and development landscape, inviting contributions from various sectors. Ultimately, this initiative represents a significant step forward in making powerful AI tools more accessible to a wider audience.

QwQ-Max-Preview represents an advanced AI model built on the Qwen2.5-Max architecture, designed to demonstrate exceptional abilities in areas such as intricate reasoning, mathematical challenges, programming tasks, and agent-based activities. This preview highlights its improved functionalities across various general-domain applications, showcasing a strong capability to handle complex workflows effectively. Set to be launched as open-source software under the Apache 2.0 license, QwQ-Max-Preview is expected to feature substantial enhancements and refinements in its final version. In addition to its technical advancements, the model plays a vital role in fostering a more inclusive AI landscape, which is further supported by the upcoming release of the Qwen Chat application and streamlined model options like QwQ-32B, aimed at developers seeking local deployment alternatives. This initiative not only enhances accessibility for a broader audience but also stimulates creativity and progress within the AI community, ensuring that diverse voices can contribute to the field's evolution. The commitment to open-source principles is likely to inspire further exploration and collaboration among developers.

Laguna M.1 is recognized as Poolside's premier model for agentic coding, meticulously designed in-house to optimize software development processes. This sophisticated model incorporates 225 billion parameters and employs a Mixture of Experts architecture with 23 billion parameters activated, all trained on a colossal dataset of 30 trillion tokens using a network of 6,144 NVIDIA H200 GPUs. Poolside committed to developing Laguna M.1 from the ground up, utilizing proprietary data, a specialized training codebase, and an asynchronous on-policy reinforcement learning strategy within its agent framework, all specifically oriented towards agentic coding applications. The model's architecture is crafted to deliver top-tier performance within Poolside's coding agent, empowering it to adeptly reason through programming tasks, engage with an array of tools, modify code, run tests, and support extensive autonomous development sessions. Tailored for developers and teams facing complex coding obstacles, Laguna M.1 boasts enhanced capabilities in reasoning, understanding architecture, managing terminal actions, and executing multi-step processes, far exceeding the abilities of lighter models. Overall, its comprehensive feature set establishes it as an indispensable tool for professionals immersed in high-stakes software projects, making it a vital component in the landscape of agentic coding solutions.

DeepSeek-V4.1-Flash is an exceptionally efficient and versatile AI model designed for tackling complex tasks in programming, creativity, agentic functions, and spatial reasoning. Building upon the foundation laid by DeepSeek-V4-Flash, this iteration emphasizes swift output generation while maintaining strong performance on challenging tasks, achieving an impressive rate of over 400 tokens per second, with peak performance reaching around 427 tokens per second in testing environments. The model is adept at managing advanced programming assignments, creating immersive 3D worlds, producing voxel-based projects, and interpreting spatially complex scenes and simulations. Remarkable demonstrations highlight its adaptability through various environments, such as Minecraft-inspired landscapes, traditional Chinese gardens, racing circuits, dungeon explorations, and exploded camera perspectives, showcasing the fusion of coding abilities and spatial understanding. Its sophisticated features make it an excellent option for rapid prototyping, game development, 3D modeling, architectural design, academic inquiries, and a wide array of other technical or artistic projects where quick iteration is essential. Moreover, the model’s innovative functionalities enable it to seamlessly adjust to various project demands, significantly increasing its effectiveness across numerous fields and enhancing collaboration between creative and technical disciplines.

GLM-5 is Z.ai’s most advanced open-source model to date, purpose-built for complex systems engineering, long-horizon planning, and autonomous agent workflows. Building on the foundation of GLM-4.5, it dramatically scales both total parameters and pre-training data while increasing active parameter efficiency. The integration of DeepSeek Sparse Attention allows GLM-5 to maintain strong long-context reasoning capabilities while reducing deployment costs. To improve post-training performance, Z.ai developed slime, an asynchronous reinforcement learning infrastructure that significantly boosts training throughput and iteration speed. As a result, GLM-5 achieves top-tier performance among open-source models across reasoning, coding, and general agent benchmarks. It demonstrates exceptional strength in long-term operational simulations, including leading results on Vending Bench 2, where it manages a year-long simulated business with strong financial outcomes. In coding evaluations such as SWE-bench and Terminal-Bench 2.0, GLM-5 delivers competitive results that narrow the gap with proprietary frontier systems. The model is fully open-sourced under the MIT License and available through Hugging Face, ModelScope, and Z.ai’s developer platforms. Developers can deploy GLM-5 locally using inference frameworks like vLLM and SGLang, including support for non-NVIDIA hardware through optimization and quantization techniques. Through Z.ai, users can access both Chat Mode for fast interactions and Agent Mode for tool-augmented, multi-step task execution. GLM-5 also enables structured document generation, producing ready-to-use .docx, .pdf, and .xlsx files for business and academic workflows. With compatibility across coding agents and cross-application automation frameworks, GLM-5 moves foundation models from conversational assistants toward full-scale work engines.

Qwen3.8-2.4T-A95B emerges as the largest open model in the Qwen3.8 series, presenting advanced Qwen-Max-class capabilities in a format that is accessible to the public. Built on the robust foundation of Qwen3.5, this model offers marked improvements in performance across various domains, including coding, professional applications, research, and complex, extended agentic tasks, underscoring its ability to reliably execute intricate, multi-step workflows to completion. With its innovative mixture-of-experts architecture, it features a remarkable total of 2.4 trillion parameters, of which 95 billion are activated, utilizing 512 experts and allowing for simultaneous engagement of 10 routed experts alongside one shared expert. The model supports a native context length of 262,144 tokens, extendable to about 1.01 million tokens, thereby enabling considerable adaptability for diverse applications. Additionally, enhancements in agent execution, such as superior autonomous planning and improved responsiveness to environmental cues, enhance its overall efficiency. Its extensive compatibility with popular agent frameworks and development tools further aids in smooth integration into current systems, making it an appealing option for both developers and researchers. This versatility is particularly beneficial for those seeking to leverage advanced AI capabilities in their projects.

Inkling-Small is an efficient multimodal AI model built to deliver strong reasoning and coding performance at a fraction of Inkling’s size. It is a Mixture-of-Experts transformer with 276 billion total parameters and 12 billion active parameters. The model was trained on NVIDIA GB300 NVL72 systems and is designed to combine high capability with more efficient inference. Inkling-Small supports native reasoning across text, images, and audio, allowing it to work across multimodal tasks without relying on separate encoders. Its context window supports up to one million tokens, making it useful for long-form reasoning, large-scale code understanding, document analysis, and agentic workflows. Users can adjust reasoning effort from minimal to extra high depending on whether they need faster responses or deeper computation. The model’s training process includes improved pre-training data, post-training with on-policy distillation from Inkling, and extended agentic coding reinforcement learning. These techniques helped Inkling-Small outperform its larger counterpart on reasoning and coding benchmarks. The model performs well in coding and tool-use harnesses and exceeds 80% on SWE-bench Verified. Its encoder-free architecture processes audio as dMel spectrograms and images as 40-by-40-pixel patches alongside text tokens. By combining efficient MoE design, one-million-token context, adjustable reasoning effort, multimodal processing, coding strength, and tool-use performance, Inkling-Small is designed for developers and teams that need capable AI with lower active compute requirements.

Qwen3.6-Plus is a cutting-edge AI model developed by Alibaba Cloud, designed to enable real-world intelligent agents, advanced coding workflows, and multimodal reasoning. It represents a major evolution in the Qwen series, offering enhanced performance across coding, reasoning, and tool-based tasks. With a default 1 million token context window, the model can process extremely large inputs and maintain context across long interactions. It excels in agentic coding, supporting tasks such as debugging, terminal operations, and large-scale repository management. The model integrates reasoning, memory, and execution capabilities, allowing it to function as a highly autonomous and reliable AI agent. Qwen3.6-Plus also features strong multimodal capabilities, enabling it to analyze images, videos, documents, and UI elements for deeper understanding and action. It supports real-world applications such as workflow automation, visual reasoning, and interactive task execution. Developers can access the model via API and integrate it with tools like OpenClaw, Qwen Code, and other coding assistants. Features like preserved reasoning context improve performance in complex, multi-step tasks and reduce redundant processing. The model is optimized for enterprise use, offering stability, scalability, and high accuracy across diverse domains. It also supports multilingual environments, making it suitable for global applications. Overall, Qwen3.6-Plus provides a powerful foundation for building next-generation AI agents capable of perception, reasoning, and action.