Alternatives to Leanstral 1.5
Compare Leanstral 1.5 alternatives for your business or organization using the curated list below. SourceForge ranks the best alternatives to Leanstral 1.5 in 2026. Compare features, ratings, user reviews, pricing, and more from Leanstral 1.5 competitors and alternatives in order to make an informed decision for your business.
-
1
TrustInSoft Analyzer
TrustInSoft
TrustInSoft Analyzer is a C/C++/Rust source code analyzer powered by formal methods, mathematical & logical reasonings that allow for exhaustive analysis of source code. This analysis can be run without false positives or false negatives, so that every real bug in the code is found. Developers receive several benefits: a user-friendly graphical interface that directs developers to the root cause of bugs, and instant utility to expand the coverage of their existing tests. Unlike traditional source code analysis tools, TrustInSoft’s solution is not only the most comprehensive approach on the market but is also progressive, instantly deployable by developers, even if they lack experience with formal methods, from exhaustive analysis up to a functional proof that the software developed meets specifications. Companies who use TrustInSoft Analyzer reduce their verification costs by 4, efforts in bug detection by 40, and obtain an irrefutable proof that their software is safe and secure. -
2
Leanstral
Mistral AI
Leanstral is an open-source code agent developed by Mistral AI specifically designed to work with the Lean 4 proof assistant. The model focuses on generating code while also formally verifying its correctness against strict mathematical or software specifications. Unlike traditional coding assistants, Leanstral integrates directly with formal proof systems to ensure that generated code satisfies defined logical requirements. Its architecture is optimized for proof engineering tasks and operates efficiently with sparse model parameters. Leanstral is released under the Apache 2.0 license, making it freely accessible for developers, researchers, and organizations to use and customize. The model is designed to operate within real-world formal repositories rather than isolated problem environments. By combining code generation with formal verification, Leanstral aims to reduce the need for manual human review in complex software and mathematical development.Starting Price: Free -
3
Harmonic Aristotle
Harmonic
Aristotle is the first AI model built from the ground up as a Mathematical Superintelligence (MSI), designed to deliver provably correct solutions to complex quantitative problems without hallucinations. When prompted with natural‑language math questions, it formalizes them in Lean 4, solves them via formally verified proofs, and returns both the proof and a natural‑language explanation. Unlike conventional language models that rely on probabilistic outputs, Aristotle’s MSI architecture replaces guesswork with provable logic, transparently flagging any errors or inconsistencies. The AI is accessible through a web interface and a developer API, enabling researchers to integrate its rigorous reasoning into workflows across fields such as theoretical physics, engineering, and computer science. -
4
DeepSWE
Agentica Project
DeepSWE is a fully open source, state-of-the-art coding agent built on top of the Qwen3-32B foundation model and trained exclusively via reinforcement learning (RL), without supervised finetuning or distillation from proprietary models. It is developed using rLLM, Agentica’s open source RL framework for language agents. DeepSWE operates as an agent; it interacts with a simulated development environment (via the R2E-Gym environment) using a suite of tools (file editor, search, shell-execution, submit/finish), enabling it to navigate codebases, edit multiple files, compile/run tests, and iteratively produce patches or complete engineering tasks. DeepSWE exhibits emergent behaviors beyond simple code generation; when presented with bugs or feature requests, the agent reasons about edge cases, seeks existing tests in the repository, proposes patches, writes extra tests for regressions, and dynamically adjusts its “thinking” effort.Starting Price: Free -
5
SWE-1.5
Cognition
SWE-1.5 is the latest agent-model release by Cognition, purpose-built for software engineering and characterized by a “frontier-size” architecture comprising hundreds of billions of parameters and optimized end-to-end (model, inference engine, and agent harness) for both speed and intelligence. It achieves near-state-of-the-art coding performance and sets a new benchmark in latency, delivering inference speeds up to 950 tokens/second, roughly six times faster than its predecessor Haiku 4.5 and thirteen times faster than Sonnet 4.5. The model was trained using extensive reinforcement learning in realistic coding-agent environments with multi-turn workflows, unit tests, quality rubrics, and browser-based agentic execution; it also benefits from tightly integrated software tooling and high-throughput hardware (including thousands of GB200 NVL72 chips and a custom hypervisor infrastructure). -
6
Command R+
Cohere AI
Command R+ is Cohere's newest large language model, optimized for conversational interaction and long-context tasks. It aims at being extremely performant, enabling companies to move beyond proof of concept and into production. We recommend using Command R+ for those workflows that lean on complex RAG functionality and multi-step tool use (agents). Command R, on the other hand, is great for simpler retrieval augmented generation (RAG) and single-step tool use tasks, as well as applications where price is a major consideration.Starting Price: Free -
7
Laguna XS.2
Poolside
Laguna XS.2 is Poolside’s open-weight agentic coding model, built as the lightest and fastest model in the Laguna family. It is a 33B total-parameter Mixture of Experts model with 3B activated parameters, trained completely in-house on 30T tokens. As Poolside’s newest generation model open to the community, Laguna XS.2 is a second-generation architecture and the company’s first open-weight model, built on the lessons learned from training Laguna M.1 across synthetic data and reinforcement learning. The model is designed for agentic coding workflows, where it can code, act, iterate quickly, and perform best inside Poolside’s coding agent. Laguna XS.2 is positioned as a strong model for rapid agentic iteration, especially for developers and teams that need a compact, efficient coding model rather than a heavier frontier system. It is released under an Apache 2.0 license, allowing the community to evaluate, fine-tune, quantize, serve, and build on the weights.Starting Price: Free -
8
Data Theorem
Data Theorem
Inventory your apps, APIs, and shadow assets across your global, multi-cloud environment. Establish custom policies for different types of asset groups, automate attack tools, and assess vulnerabilities. Fix security issues before going into production, making sure application and cloud data is compliant. Auto-remediation of vulnerabilities with rollback options to stop leaky data. Good security finds problems fast, but great security makes problems disappear. Data Theorem strives to make great products that automate the most challenging areas of modern application security. The core of Data Theorem is its Analyzer Engine. Utilize the Data Theorem analyzer engine & proprietary attack tools to hack and exploit application weaknesses continuously. Data Theorem has built the top open source SDK called TrustKit, used by thousands of developers. Our technology ecosystem continues to grow so that customers can continue to secure their entire Appsec stack with ease. -
9
Qwen3-Coder
Qwen
Qwen3‑Coder is an agentic code model available in multiple sizes, led by the 480B‑parameter Mixture‑of‑Experts variant (35B active) that natively supports 256K‑token contexts (extendable to 1M) and achieves state‑of‑the‑art results comparable to Claude Sonnet 4. Pre‑training on 7.5T tokens (70 % code) and synthetic data cleaned via Qwen2.5‑Coder optimized both coding proficiency and general abilities, while post‑training employs large‑scale, execution‑driven reinforcement learning, scaling test‑case generation for diverse coding challenges, and long‑horizon RL across 20,000 parallel environments to excel on multi‑turn software‑engineering benchmarks like SWE‑Bench Verified without test‑time scaling. Alongside the model, the open source Qwen Code CLI (forked from Gemini Code) unleashes Qwen3‑Coder in agentic workflows with customized prompts, function calling protocols, and seamless integration with Node.js, OpenAI SDKs, and environment variables.Starting Price: Free -
10
Hunyuan Motion 1.0
Tencent Hunyuan
Hunyuan Motion (also known as HY-Motion 1.0) is a state-of-the-art text-to-3D motion generation AI model that uses a billion-parameter Diffusion Transformer with flow matching to turn natural language prompts into high-quality, skeleton-based 3D character animation in seconds. It understands descriptive text in English and Chinese and produces smooth, physically plausible motion sequences that integrate seamlessly into standard 3D animation pipelines by exporting to skeleton formats such as SMPL or SMPLH and common formats like FBX or BVH for use in Blender, Unity, Unreal Engine, Maya, and other tools. The model’s three-stage training pipeline (large-scale pre-training on thousands of hours of motion data, fine-tuning on curated sequences, and reinforcement learning from human feedback) enhances its ability to follow complex instructions and generate realistic, temporally coherent motion. -
11
Theorem
Theorem Technologies
Theorem is a post-trade processing and data management solution that delivers aggregated reporting, trade workflow, matching / affirmation, and data movement, all in an online portal that requires no local install or technician to get started. Use Theorem to connect data collected from brokers, exchanges, and trading systems or your own internal systems and manage it according to your specific needs. Since there’s no heavy install, leverage the entire platform for a complete solution or simply choose what you need now. The moment a trade is executed, Theorem ensures accurate post-trade processing by allocating trades and performing trade matching against executing and clearing counterparties. Theorem harmonizes data, analytics, and other insights across brokers, accounts, and portfolios, offering you access to actionable insights while providing your stakeholders with advanced analytics and easily understood risk reports.Starting Price: $500.00/month -
12
Olmo 3
Ai2
Olmo 3 is a fully open model family spanning 7 billion and 32 billion parameter variants that delivers not only high-performing base, reasoning, instruction, and reinforcement-learning models, but also exposure of the entire model flow, including raw training data, intermediate checkpoints, training code, long-context support (65,536 token window), and provenance tooling. Starting with the Dolma 3 dataset (≈9 trillion tokens) and its disciplined mix of web text, scientific PDFs, code, and long-form documents, the pre-training, mid-training, and long-context phases shape the base models, which are then post-trained via supervised fine-tuning, direct preference optimisation, and RL with verifiable rewards to yield the Think and Instruct variants. The 32 B Think model is described as the strongest fully open reasoning model to date, competitively close to closed-weight peers in math, code, and complex reasoning.Starting Price: Free -
13
Ring 2.6
Ant Group
Ring is a trillion-parameter thinking model from Ant Group, designed for real-world Agent workflows. It uses the same Mixture of Experts architecture as Ling, activating about 63B parameters per inference, and focuses on coding agents, tool use, multi-tool collaboration, engineering development, research analysis, and long-horizon task execution. Rather than only pursuing “smarter” results, Ring is built to consistently complete complex tasks at reasonable cost, balancing quality, speed, and execution efficiency in production environments. Ring-2.6-1T introduces an adjustable Reasoning Effort mechanism with high and xhigh reasoning intensity levels, using adaptive reasoning budget allocation based on task complexity. High mode is designed for high-frequency Agent workflows, lower token cost, faster multi-step execution, multi-turn interaction, tool collaboration, and task decomposition.Starting Price: $0.0028 per 1M tokens -
14
Phi-4-reasoning
Microsoft
Phi-4-reasoning is a 14-billion parameter transformer-based language model optimized for complex reasoning tasks, including math, coding, algorithmic problem solving, and planning. Trained via supervised fine-tuning of Phi-4 on carefully curated "teachable" prompts and reasoning demonstrations generated using o3-mini, it generates detailed reasoning chains that effectively leverage inference-time compute. Phi-4-reasoning incorporates outcome-based reinforcement learning to produce longer reasoning traces. It outperforms significantly larger open-weight models such as DeepSeek-R1-Distill-Llama-70B and approaches the performance levels of the full DeepSeek-R1 model across a wide range of reasoning tasks. Phi-4-reasoning is designed for environments with constrained computing or latency. Fine-tuned with synthetic data generated by DeepSeek-R1, it provides high-quality, step-by-step problem solving. -
15
Sarvam-M
Sarvam
Sarvam-M is a multilingual, hybrid-reasoning large language model designed to deliver strong performance across Indian languages, mathematical reasoning, and programming tasks within a single, efficient system. Built on top of Mistral-Small, it is a 24-billion-parameter text-only model that has been enhanced through supervised fine-tuning, reinforcement learning with verifiable rewards, and inference optimizations to improve both accuracy and efficiency. The model is specifically trained to handle more than ten major Indic languages, supporting native scripts, romanized text, and code-mixed inputs, enabling seamless multilingual communication across diverse linguistic contexts. Sarvam-M introduces a hybrid reasoning approach that allows it to switch between “thinking” mode for complex tasks like math, logic, and coding, and faster response mode for everyday interactions, balancing performance and speed. -
16
Laguna M.1
Poolside
Laguna M.1 is Poolside’s most capable model for agentic coding, built and trained in-house for software development workflows. It is a 225B total-parameter Mixture of Experts model with 23B activated parameters, trained completely in-house on 30T tokens using 6,144 interconnected NVIDIA H200 GPUs. Poolside trained Laguna M.1 from scratch with its own data work, training codebase, and async on-policy reinforcement learning in its agent harness, all with agentic coding in mind. The model is designed to perform at its best inside Poolside’s coding agent, where it can reason through software tasks, interact with tools, edit code, run tests, and support longer autonomous development sessions. Laguna M.1 is built for developers and teams working on complex coding tasks that require stronger reasoning, architectural understanding, terminal use, and multi-step execution than lightweight models can provide.Starting Price: Free -
17
Sparrow
DeepMind
Sparrow is a research model and proof of concept, designed with the goal of training dialogue agents to be more helpful, correct, and harmless. By learning these qualities in a general dialogue setting, Sparrow advances our understanding of how we can train agents to be safer and more useful – and ultimately, to help build safer and more useful artificial general intelligence (AGI). Sparrow is not yet available for public use. Training a conversational AI is an especially challenging problem because it’s difficult to pinpoint what makes a dialogue successful. To address this problem, we turn to a form of reinforcement learning (RL) based on people's feedback, using the study participants’ preference feedback to train a model of how useful an answer is. To get this data, we show our participants multiple model answers to the same question and ask them which answer they like the most. -
18
ChatGLM
Zhipu AI
ChatGLM-6B is an open-source, Chinese-English bilingual dialogue language model based on the General Language Model (GLM) architecture with 6.2 billion parameters. Combined with model quantization technology, users can deploy locally on consumer-grade graphics cards (only 6GB of video memory is required at the INT4 quantization level). ChatGLM-6B uses technology similar to ChatGPT, optimized for Chinese Q&A and dialogue. After about 1T identifiers of Chinese and English bilingual training, supplemented by supervision and fine-tuning, feedback self-help, human feedback reinforcement learning and other technologies, ChatGLM-6B with 6.2 billion parameters has been able to generate answers that are quite in line with human preferences.Starting Price: Free -
19
KAT-Coder-Pro V2
StreamLake
KAT-Coder is an agentic AI coding system designed to go beyond traditional autocomplete tools by enabling end-to-end software development workflows driven by reasoning, planning, and execution. It is positioned as a flagship coding model within the KAT ecosystem, built specifically for “agentic coding,” where the model does not just generate snippets but can diagnose issues, propose fixes, run tests, and iterate across multiple files as part of a continuous development loop. It integrates directly with developer environments through API endpoints and proxy layers compatible with tools like Claude Code, allowing seamless use inside existing IDE workflows without changing the interface developers are already familiar with. KAT-Coder is trained using a multi-stage pipeline that includes supervised fine-tuning and large-scale reinforcement learning, enabling it to understand programming context, and reason over complex tasks.Starting Price: $0.30 per month -
20
Prolog
Prolog
Prolog is a logic programming language associated with artificial intelligence and computational linguistics. Prolog has its roots in first-order logic, a formal logic, and unlike many other programming languages, Prolog is intended primarily as a declarative programming language, the program logic is expressed in terms of relations, represented as facts and rules. A computation is initiated by running a query over these relations. Prolog was one of the first logic programming languages and remains the most popular such language today, with several free and commercial implementations available. The language has been used for theorem proving, expert systems, term rewriting, type systems, and automated planning, as well as its original intended field of use, natural language processing. Modern Prolog environments support the creation of graphical user interfaces, as well as administrative and networked applications. -
21
GLM-5
Zhipu AI
GLM-5 is Z.ai’s latest large language model built for complex systems engineering and long-horizon agentic tasks. It scales significantly beyond GLM-4.5, increasing total parameters and training data while integrating DeepSeek Sparse Attention to reduce deployment costs without sacrificing long-context capacity. The model combines enhanced pre-training with a new asynchronous reinforcement learning infrastructure called slime, improving training efficiency and post-training refinement. GLM-5 achieves best-in-class performance among open-source models across reasoning, coding, and agent benchmarks, narrowing the gap with leading frontier models. It ranks highly on evaluations such as Vending Bench 2, demonstrating strong long-term planning and operational capabilities. The model is open-sourced under the MIT License.Starting Price: Free -
22
FLUX.1 Krea
Krea
FLUX.1 Krea is an open source, guidance-distilled 12 billion-parameter diffusion transformer released by Krea in collaboration with Black Forest Labs, engineered to deliver superior aesthetic control and photorealism while eschewing the generic “AI look.” Fully compatible with the FLUX.1-dev ecosystem, it starts from a raw, untainted base model (flux-dev-raw) rich in world knowledge and employs a two-phase post-training pipeline, supervised fine-tuning on a hand-curated mix of high-quality and synthetic samples, followed by reinforcement learning from human feedback using opinionated preference data, to bias outputs toward a distinct style. By leveraging negative prompts during pre-training, custom loss functions for classifier-free guidance, and targeted preference labels, it achieves significant quality improvements with under one million examples, all without extensive prompting or additional LoRA modules.Starting Price: Free -
23
Grok 4.1 Fast
xAI
Grok 4.1 Fast is an xAI model designed to deliver advanced tool-calling capabilities with a massive 2-million-token context window. It excels at complex real-world tasks such as customer support, finance, troubleshooting, and dynamic agent workflows. The model pairs seamlessly with the new Agent Tools API, which enables real-time web search, X search, file retrieval, and secure code execution. This combination gives developers the power to build fully autonomous, production-grade agents that plan, reason, and use tools effectively. Grok 4.1 Fast is trained with long-horizon reinforcement learning, ensuring stable multi-turn accuracy even across extremely long prompts. With its speed, cost-efficiency, and high benchmark scores, it sets a new standard for scalable enterprise-grade AI agents. -
24
Kimi K2
Moonshot AI
Kimi K2 is a state-of-the-art open source large language model series built on a mixture-of-experts (MoE) architecture, featuring 1 trillion total parameters and 32 billion activated parameters for task-specific efficiency. Trained with the Muon optimizer on over 15.5 trillion tokens and stabilized by MuonClip’s attention-logit clamping, it delivers exceptional performance in frontier knowledge, reasoning, mathematics, coding, and general agentic workflows. Moonshot AI provides two variants, Kimi-K2-Base for research-level fine-tuning and Kimi-K2-Instruct pre-trained for immediate chat and tool-driven interactions, enabling both custom development and drop-in agentic capabilities. Benchmarks show it outperforms leading open source peers and rivals top proprietary models in coding tasks and complex task breakdowns, while its 128 K-token context length, tool-calling API compatibility, and support for industry-standard inference engines.Starting Price: Free -
25
MAI-Code-1-Flash
Microsoft AI
MAI-Code-1-Flash is a Microsoft coding model built for fast, efficient assistance in everyday developer workflows. Built end-to-end by Microsoft using clean and appropriately licensed data, the model is rolling out to GitHub Copilot individual users in Visual Studio Code through the model picker and the default Auto picker. It is designed around the goal of delivering high-quality coding help with better efficiency, helping engineering teams write better code faster through a lightweight, agentic model integrated into GitHub Copilot and VS Code. MAI-Code-1-Flash was trained directly with GitHub Copilot production harnesses, allowing it to interact with surrounding tools and systems in real developer environments rather than being optimized only for static benchmarks. It supports agentic coding, strong instruction-following across single-turn and multi-turn scenarios, repository question answering, refactoring, telemetry-grounded tasks, and adaptive thinking. -
26
ZenCtrl
Fotographer AI
ZenCtrl is an open source AI image generation toolkit developed by Fotographer AI, designed to produce high-quality, multi-view, and diverse-scene outputs from a single image without any training. It enables precise regeneration of objects and subjects from any angle and background, offering real-time element regeneration that provides both stability and flexibility in creative workflows. ZenCtrl allows users to regenerate subjects from any angle, swap backgrounds or clothing with just a click, and start generating results immediately without the need for additional training. By leveraging advanced image processing techniques, it ensures high accuracy without the need for extensive training data. The model's architecture is composed of lightweight sub-models, each fine-tuned on task-specific data to excel at a single job, resulting in a lean system that delivers sharper, more controllable results.Starting Price: Free -
27
Orpheus TTS
Canopy Labs
Canopy Labs has introduced Orpheus, a family of state-of-the-art speech large language models (LLMs) designed for human-level speech generation. These models are built on the Llama-3 architecture and are trained on over 100,000 hours of English speech data, enabling them to produce natural intonation, emotion, and rhythm that surpasses current state-of-the-art closed source models. Orpheus supports zero-shot voice cloning, allowing users to replicate voices without prior fine-tuning, and offers guided emotion and intonation control through simple tags. The models achieve low latency, with approximately 200ms streaming latency for real-time applications, reducible to around 100ms with input streaming. Canopy Labs has released both pre-trained and fine-tuned 3B-parameter models under the permissive Apache 2.0 license, with plans to release smaller models of 1B, 400M, and 150M parameters for use on resource-constrained devices. -
28
Retrocausal
Retrocausal
Empower your operators, engineers, and managers to dramatically boost the quality and productivity of your manual processes. Create digital mistake-proofing mechanisms for a variety of assembly and packing processes. Assembly Copilot tracks individual steps of an assembly process and offers audible and visual alerts to help associates avoid mistakes. Copilot also provides native support for Signal Towers and Projectors to reinforce its alerts. Assembly Copilot measures cycle times, step-level statistics, variations, and non-value-added activities. This highlights process variability and helps industrial and lean engineers balance lines. Copilot helps manufacturers extract more productivity out of their processes. Assembly Copilot extends the typical notion of Total Quality Management (TQM) beyond the traceability of parts, to every assembly step. -
29
GPT‑5.4‑Cyber
OpenAI
GPT-5.4-Cyber is a specialized, cyber-permissive variant of GPT-5.4 designed specifically to support defensive cybersecurity workflows, enabling security professionals to analyze, detect, and remediate vulnerabilities more effectively. It is fine-tuned to lower the refusal boundary for legitimate security tasks, allowing deeper engagement with activities such as vulnerability research, exploit analysis, and secure code evaluation that are typically restricted in general-purpose models. A key capability includes binary reverse engineering, which allows the model to analyze compiled software without access to source code to identify malware potential, weaknesses, and overall system robustness. Integrated within OpenAI’s Trusted Access for Cyber (TAC) program, the model is distributed through a tiered access system that requires identity verification and progressive trust levels, ensuring that only vetted defenders, researchers, and organizations can access its most advanced features.Starting Price: Free -
30
MiMo-V2-Flash
Xiaomi Technology
MiMo-V2-Flash is an open weight large language model developed by Xiaomi based on a Mixture-of-Experts (MoE) architecture that blends high performance with inference efficiency. It has 309 billion total parameters but activates only 15 billion active parameters per inference, letting it balance reasoning quality and computational efficiency while supporting extremely long context handling, for tasks like long-document understanding, code generation, and multi-step agent workflows. It incorporates a hybrid attention mechanism that interleaves sliding-window and global attention layers to reduce memory usage and maintain long-range comprehension, and it uses a Multi-Token Prediction (MTP) design that accelerates inference by processing batches of tokens in parallel. MiMo-V2-Flash delivers very fast generation speeds (up to ~150 tokens/second) and is optimized for agentic applications requiring sustained reasoning and multi-turn interactions.Starting Price: Free -
31
Qwen Code
Qwen
Qwen3‑Coder is an agentic code model available in multiple sizes, led by the 480B‑parameter Mixture‑of‑Experts variant (35B active) that natively supports 256K‑token contexts (extendable to 1M) and achieves state‑of‑the‑art results on Agentic Coding, Browser‑Use, and Tool‑Use tasks comparable to Claude Sonnet 4. Pre‑training on 7.5T tokens (70 % code) and synthetic data cleaned via Qwen2.5‑Coder optimized both coding proficiency and general abilities, while post‑training employs large‑scale, execution‑driven reinforcement learning and long‑horizon RL across 20,000 parallel environments to excel on multi‑turn software‑engineering benchmarks like SWE‑Bench Verified without test‑time scaling. Alongside the model, the open source Qwen Code CLI (forked from Gemini Code) unleashes Qwen3‑Coder in agentic workflows with customized prompts, function calling protocols, and seamless integration with Node.js, OpenAI SDKs, and more.Starting Price: Free -
32
Tülu 3
Ai2
Tülu 3 is an advanced instruction-following language model developed by the Allen Institute for AI (Ai2), designed to enhance capabilities in areas such as knowledge, reasoning, mathematics, coding, and safety. Built upon the Llama 3 Base, Tülu 3 employs a comprehensive four-stage post-training process: meticulous prompt curation and synthesis, supervised fine-tuning on a diverse set of prompts and completions, preference tuning using both off- and on-policy data, and a novel reinforcement learning approach to bolster specific skills with verifiable rewards. This open-source model distinguishes itself by providing full transparency, including access to training data, code, and evaluation tools, thereby closing the performance gap between open and proprietary fine-tuning methods. Evaluations indicate that Tülu 3 outperforms other open-weight models of similar size, such as Llama 3.1-Instruct and Qwen2.5-Instruct, across various benchmarks.Starting Price: Free -
33
Bookdown
Bookdown
Write HTML, PDF, ePub, and Kindle books with R Markdown. The bookdown package is an open-source R package that facilitates writing books and long-form articles/reports with R Markdown. Generate printer-ready books and ebooks from R Markdown documents. A markup language easier to learn than LaTeX, and to write elements such as section headers, lists, quotes, figures, tables, and citations. Multiple choices of output formats: PDF, LaTeX, HTML, EPUB, and Word. Possibility of including dynamic graphics and interactive applications (HTML widgets and Shiny apps). Support a wide range of languages: R, C/C++, Python, Fortran, Julia, Shell scripts, and SQL, etc. LaTeX equations, theorems, and proofs work for all output formats. Can be published to GitHub, bookdown.org, and any web servers. Integrated with the RStudio IDE. One-click publishing to https://bookdown.org. -
34
Nemotron 3 Ultra
NVIDIA
Nemotron 3 Nano is a compact, open large language model in NVIDIA’s Nemotron 3 family, designed for efficient agentic reasoning, conversational AI, and coding tasks. It uses a hybrid Mixture-of-Experts Mamba-Transformer architecture that activates only a small subset of parameters per token, enabling low-latency inference while maintaining strong accuracy and reasoning performance. It has approximately 31.6 billion total parameters with around 3.2 billion active (3.6 billion including embeddings), allowing it to achieve higher accuracy than previous Nemotron 2 Nano while using less computation per forward pass. Nemotron 3 Nano supports long-context processing of up to one million tokens, enabling it to handle large documents, multi-step workflows, and extended reasoning chains in a single pass. It is designed for high-throughput, real-time execution, excelling in multi-turn conversations, tool calling, and agent-based workflows where tasks require planning, reasoning, and more. -
35
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. -
36
DeepCoder
Agentica Project
DeepCoder is a fully open source code-reasoning and generation model released by Agentica Project in collaboration with Together AI. It is fine-tuned from DeepSeek-R1-Distilled-Qwen-14B using distributed reinforcement learning, achieving a 60.6% accuracy on LiveCodeBench (representing an 8% improvement over the base), a performance level that matches that of proprietary models such as o3-mini (2025-01-031 Low) and o1 while using only 14 billion parameters. It was trained over 2.5 weeks on 32 H100 GPUs with a curated dataset of roughly 24,000 coding problems drawn from verified sources (including TACO-Verified, PrimeIntellect SYNTHETIC-1, and LiveCodeBench submissions), each problem requiring a verifiable solution and at least five unit tests to ensure reliability for RL training. To handle long-range context, DeepCoder employs techniques such as iterative context lengthening and overlong filtering.Starting Price: Free -
37
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. -
38
Xgen-small
Salesforce
Xgen-small is an enterprise-ready compact language model developed by Salesforce AI Research, designed to deliver long-context performance at a predictable, low cost. It combines domain-focused data curation, scalable pre-training, length extension, instruction fine-tuning, and reinforcement learning to meet the complex, high-volume inference demands of modern enterprises. Unlike traditional large models, Xgen-small offers efficient processing of extensive contexts, enabling the synthesis of information from internal documentation, code repositories, research reports, and real-time data streams. With sizes optimized at 4B and 9B parameters, it provides a strategic advantage by balancing cost efficiency, privacy safeguards, and long-context understanding, making it a sustainable and predictable solution for deploying Enterprise AI at scale. -
39
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. -
40
Mistral NeMo
Mistral AI
Mistral NeMo, our new best small model. A state-of-the-art 12B model with 128k context length, and released under the Apache 2.0 license. Mistral NeMo is a 12B model built in collaboration with NVIDIA. Mistral NeMo offers a large context window of up to 128k tokens. Its reasoning, world knowledge, and coding accuracy are state-of-the-art in its size category. As it relies on standard architecture, Mistral NeMo is easy to use and a drop-in replacement in any system using Mistral 7B. We have released pre-trained base and instruction-tuned checkpoints under the Apache 2.0 license to promote adoption for researchers and enterprises. Mistral NeMo was trained with quantization awareness, enabling FP8 inference without any performance loss. The model is designed for global, multilingual applications. It is trained on function calling and has a large context window. Compared to Mistral 7B, it is much better at following precise instructions, reasoning, and handling multi-turn conversations.Starting Price: Free -
41
Hermes 3
Nous Research
Experiment, and push the boundaries of individual alignment, artificial consciousness, open-source software, and decentralization, in ways that monolithic companies and governments are too afraid to try. Hermes 3 contains advanced long-term context retention and multi-turn conversation capability, complex roleplaying and internal monologue abilities, and enhanced agentic function-calling. Our training data aggressively encourages the model to follow the system and instruction prompts exactly and in an adaptive manner. Hermes 3 was created by fine-tuning Llama 3.1 8B, 70B, and 405B, and training on a dataset of primarily synthetically generated responses. The model boasts comparable and superior performance to Llama 3.1 while unlocking deeper capabilities in reasoning and creativity. Hermes 3 is a series of instruct and tool-use models with strong reasoning and creative abilities.Starting Price: Free -
42
ReinforceNow
ReinforceNow
ReinforceNow is an end-to-end platform for continual learning with AI agents, built to help teams deploy, train, and repeat. It lets developers build AI agents and continuously train them on production traffic, or let Claude Code help set it up automatically. It handles reinforcement learning infrastructure, experiment orchestration, agent versioning, GPU training logic, and telemetry, so teams can focus on agent logic, data collection, and rewards. ReinforceNow supports fast LLM fine-tuning with LoRA, high-throughput training, and wide model support for open source models like Qwen, DeepSeek, and GPT-OSS. It provides advanced telemetry to evaluate, monitor, and iterate on AI agent LLM applications, with traces, rewards, experiment metrics, and training observability. Teams can train on long-horizon tasks with 32k to 1 million context size, build vertical agents for multi-turn and long-running tasks, and use rich tooling for reinforcement learning workflows. -
43
Mistral 7B
Mistral AI
Mistral 7B is a 7.3-billion-parameter language model that outperforms larger models like Llama 2 13B across various benchmarks. It employs Grouped-Query Attention (GQA) for faster inference and Sliding Window Attention (SWA) to efficiently handle longer sequences. Released under the Apache 2.0 license, Mistral 7B is accessible for deployment across diverse platforms, including local environments and major cloud services. Additionally, a fine-tuned version, Mistral 7B Instruct, demonstrates enhanced performance in instruction-following tasks, surpassing models like Llama 2 13B Chat.Starting Price: Free -
44
ERNIE 5.1
Baidu
ERNIE 5.1 is Baidu’s latest large language model designed to deliver advanced reasoning, agentic AI capabilities, creative writing, and world knowledge performance while operating with significantly improved efficiency. The model builds on the foundation of ERNIE 5.0 while reducing total parameters and training costs, allowing it to achieve flagship-level intelligence at a fraction of the computational expense of comparable models. ERNIE 5.1 performs strongly across international benchmarks for reasoning, search, knowledge, and agentic tasks, ranking among the top global AI models and leading among Chinese-developed models on multiple leaderboards. The platform introduces a new fully asynchronous reinforcement learning infrastructure that improves training efficiency, scalability, and stability for complex long-horizon AI tasks. ERNIE 5.1 also features advanced creative writing capabilities. -
45
Falcon-40B
Technology Innovation Institute (TII)
Falcon-40B is a 40B parameters causal decoder-only model built by TII and trained on 1,000B tokens of RefinedWeb enhanced with curated corpora. It is made available under the Apache 2.0 license. Why use Falcon-40B? It is the best open-source model currently available. Falcon-40B outperforms LLaMA, StableLM, RedPajama, MPT, etc. See the OpenLLM Leaderboard. It features an architecture optimized for inference, with FlashAttention and multiquery. It is made available under a permissive Apache 2.0 license allowing for commercial use, without any royalties or restrictions. ⚠️ This is a raw, pretrained model, which should be further finetuned for most usecases. If you are looking for a version better suited to taking generic instructions in a chat format, we recommend taking a look at Falcon-40B-Instruct.Starting Price: Free -
46
Hy3
Tencent
Hy3 preview is Tencent Hy’s most intelligent model in the Hy series to date, built as a 295B-parameter Mixture-of-Experts model with 21B activated parameters, 3.8B MTP layer parameters, and support for up to a 256K token context window. As the first model trained on Tencent Hy’s rebuilt infrastructure, Hy3 preview is designed to improve real-world usability across complex reasoning, instruction following, context learning, coding, agent capabilities, and overall inference performance. It integrates both fast and slow thinking capabilities, allowing direct responses for simpler tasks and deeper reasoning for complex math, coding, and reasoning work. The model is built around well-rounded capabilities across long-context understanding, instruction following, tool use, and agent workflows, with evaluation focused not only on standard benchmarks but also on authentic business and development scenarios.Starting Price: Free -
47
Step 3.5 Flash
StepFun
Step 3.5 Flash is an advanced open source foundation language model engineered for frontier reasoning and agentic capabilities with exceptional efficiency, built on a sparse Mixture of Experts (MoE) architecture that selectively activates only about 11 billion of its ~196 billion parameters per token to deliver high-density intelligence and real-time responsiveness. Its 3-way Multi-Token Prediction (MTP-3) enables generation throughput in the hundreds of tokens per second for complex multi-step reasoning chains and task execution, and it supports efficient long contexts with a hybrid sliding window attention approach that reduces computational overhead across large datasets or codebases. It demonstrates robust performance on benchmarks for reasoning, coding, and agentic tasks, rivaling or exceeding many larger proprietary models, and includes a scalable reinforcement learning framework for consistent self-improvement.Starting Price: Free -
48
MAI-Thinking-1
Microsoft AI
MAI-Thinking-1 is Microsoft AI’s reasoning model, built for complex problems that matter most, with competitive reasoning and strong software engineering performance in its weight class. It is a 35B-active, approximately 1T-total-parameter sparse Mixture of Experts model, giving it a smaller inference footprint than much larger models while still matching leading models on key software engineering benchmarks. Microsoft trained MAI-Thinking-1 from the ground up on enterprise-grade, clean, commercially licensed data, without distillation from third-party models, so its capabilities are learned rather than inherited. The model is part of Microsoft AI’s Hill-Climbing Machine, a co-designed development pipeline built to make every component of model development continually and reliably improve over time. MAI-Thinking-1 is designed for agentic coding environments where models must read code, edit files, run tests, observe failures, and recover from intermediate mistakes. -
49
Qwen3-Coder-Next
Alibaba
Qwen3-Coder-Next is an open-weight language model specifically designed for coding agents and local development that delivers advanced coding reasoning, complex tool usage, and robust performance on long-horizon programming tasks with high efficiency, using a mixture-of-experts architecture that balances powerful capabilities with resource-friendly operation. It provides enhanced agentic coding abilities that help software developers, AI system builders, and automated coding workflows generate, debug, and reason about code with deep contextual understanding while recovering from execution errors, making it well-suited for autonomous coding agents and development-oriented applications. By achieving strong performance comparable to much larger parameter models while requiring fewer active parameters, Qwen3-Coder-Next enables cost-effective deployment for dynamic and complex programming workloads in research and production environments.Starting Price: Free -
50
Reka
Reka
Our enterprise-grade multimodal assistant carefully designed with privacy, security, and efficiency in mind. We train Yasa to read text, images, videos, and tabular data, with more modalities to come. Use it to generate ideas for creative tasks, get answers to basic questions, or derive insights from your internal data. Generate, train, compress, or deploy on-premise with a few simple commands. Use our proprietary algorithms to personalize our model to your data and use cases. We design proprietary algorithms involving retrieval, fine-tuning, self-supervised instruction tuning, and reinforcement learning to tune our model on your datasets.
