Search Results for "frontier"
Sort By:
Showing 23 open source projects for "frontier"
View related business solutions-
Ship Agents FasterGemini Enterprise Agent Platform lets you rapidly build, scale, govern and optimize production-ready agents grounded in your organization's data. The platform enables developers to build custom or pre-built agents for virtually any use case. New customers get $300 in free credits.
-
Veeam Data Platform v13.1 - Get Your Free TrialTry Veeam Data Platform today. Experience the unified platform that's secure by design, portable by default, and proven to recover clean, fast, and anywhere.
-
1
Kimi-K3
Open Frontier Intelligence
Kimi K3 is an open-weight, native multimodal agentic model from Moonshot AI for advanced reasoning, coding, and knowledge work. It uses a 2.8-trillion-parameter mixture-of-experts architecture while activating 104 billion parameters per token. Its design combines Kimi Delta Attention, gated multihead latent attention, Attention Residuals, and Stable LatentMoE routing. The model can process text and images and supports a context window of more than one million tokens. It is built for long... -
2
MiniMax-M2.5
State of the art LLM and coding model
...With native serving speeds of up to 100 tokens per second, it completes complex agentic tasks significantly faster than previous versions while maintaining high token efficiency. M2.5 is built to be highly cost-effective, enabling continuous deployment of powerful AI agents at a fraction of the cost of other frontier models. -
3
Kimi K2
Kimi K2 is the large language model series developed by Moonshot AI
Kimi K2 is Moonshot AI’s advanced open-source large language model built on a scalable Mixture-of-Experts (MoE) architecture that combines a trillion total parameters with a subset of ~32 billion active parameters to deliver powerful and efficient performance on diverse tasks. It was trained on an enormous corpus of over 15.5 trillion tokens to push frontier capabilities in coding, reasoning, and general agentic tasks while addressing training stability through novel optimizer and architecture design strategies. The model family includes variants like a foundational base model that researchers can fine-tune for specific use cases and an instruct-optimized variant primed for general-purpose chat and agent-style interactions, offering flexibility for both experimentation and deployment. ... -
4
VibeVoice
Open-source multi-speaker long-form text-to-speech model
VibeVoice-1.5B is Microsoft’s frontier open-source text-to-speech (TTS) model designed for generating expressive, long-form, multi-speaker conversational audio such as podcasts. Unlike traditional TTS systems, it excels in scalability, speaker consistency, and natural turn-taking for up to 90 minutes of continuous speech with as many as four distinct speakers. A key innovation is its use of continuous acoustic and semantic speech tokenizers operating at an ultra-low frame rate of 7.5 Hz, enabling high audio fidelity with efficient processing of long sequences. ... -
Go from Code to Production URL in SecondsSkip the Kubernetes configs. Cloud Run handles HTTPS, scaling, and infrastructure automatically. Two million requests free per month.
-
5
Step 3.5 Flash
Fast, Sharp & Reliable Agentic Intelligence
Step 3.5 Flash is a cutting-edge, open-source large language model developed by StepFun-AI that pushes the frontier of efficient reasoning and “agentic” intelligence in a way that makes powerful AI accessible beyond proprietary black boxes. Unlike dense models that activate all their parameters for every token, Step 3.5 Flash uses a sparse Mixture-of-Experts (MoE) architecture that selectively engages only about 11 billion of its roughly 196 billion total parameters per token, delivering high-quality reasoning and interaction at far lower compute cost and latency than traditional large models. ... -
6
VibeThinker
Diversity-driven optimization and large-model reasoning ability
VibeThinker is a compact but high-capability open-source language model released by WeiboAI (Sina AI Lab). It contains about 1.5 billion parameters, far smaller than many “frontier” models, yet it is explicitly optimized for reasoning, mathematics, and code generation tasks rather than general open-domain chat. The innovation lies in its training methodology: the team uses what they call the Spectrum-to-Signal Principle (SSP), where a first stage emphasizes diversity of reasoning paths (the “spectrum” phase) and a second stage uses reinforcement techniques (the “signal” phase) to refine toward correctness and strong reasoning. ... -
7
Step-Video-T2V
State-of-the-art (SoTA) text-to-video pre-trained model
...Its training and generation pipeline includes techniques like flow-matching, full 3D attention for temporal consistency, and fine-tuning approaches (e.g. video-based DPO) to improve fidelity and reduce artifacts. As a result, Step-Video-T2V aims to push the frontier of open-source video generation. -
8
MiniMax-M1
Open-weight, large-scale hybrid-attention reasoning model
MiniMax-M1 is presented as the world’s first open-weight, large-scale hybrid-attention reasoning model, designed to push the frontier of long-context, tool-using, and deeply “thinking” language models. It is built on the MiniMax-Text-01 foundation and keeps the same massive parameter budget, but reworks the attention and training setup for better reasoning and test-time compute scaling. Architecturally, it combines Mixture-of-Experts layers with lightning attention, enabling the model to support a native context length of 1 million tokens while using far fewer FLOPs than comparable reasoning models for very long generations. ... -
9
Kimi K3
Powerful, native multimodal AI agentic model
...Built with 2.8 trillion total parameters and a sparse mixture-of-experts architecture, it activates 104 billion parameters per token to with 2.8 trillion total parameters and a sparse mixture-of-experts architecture, it activates 104 billion parameters per token to deliver frontier-level performance more efficiently. Kimi K3 combines native text and image understanding deliver frontier-level performance more efficiently. -
99.99% Uptime for MySQL and PostgreSQL DatabasesCloud SQL Enterprise Plus delivers near-zero downtime with 35 days of point-in-time recovery. Supports MySQL, PostgreSQL, and SQL Server.
-
10
Mistral Large 3 675B Base 2512
Frontier-scale 675B multimodal base model for custom AI training
Mistral Large 3 675B Base 2512 is the foundational, pre-trained version of the Mistral Large 3 family, built as a frontier-scale multimodal Mixture-of-Experts model with 41B active parameters and a total size of 675B. It is trained from scratch using 3000 H200 GPUs, making it one of the most advanced and compute-intensive open-weight models available. As the base version, it is not fine-tuned for instruction following or reasoning, making it ideal for teams planning their own domain-specific finetuning or custom training pipelines. ... -
11
LongCat-2.0
Trillion-parameter MoE model for coding and million-token reasoning
LongCat-2.0 is Meituan’s flagship open-weight Mixture-of-Experts language model designed for frontier-scale coding, reasoning, and autonomous agent workflows. It features 1.6 trillion total parameters with approximately 48 billion activated per token, combining high capability with efficient sparse inference. The model was pretrained on more than 35 trillion tokens and trained entirely on a large-scale cluster of domestically developed AI accelerators, demonstrating stable frontier-scale training without rollback events. ... -
12
Inkling
Frontier multimodal MoE model for coding and AI agent workflows
Inkling is Thinking Machines Lab’s first open-weight flagship multimodal Mixture-of-Experts model, designed for advanced reasoning, coding, and autonomous agent workflows. It contains 975B total parameters with 41B active parameters per token, balancing frontier-level capability with efficient sparse inference. The model natively processes text, images, audio, and video within a unified architecture and supports an exceptionally large 1 million token context window for long-document reasoning, repository-scale coding, and agentic execution. Trained from scratch on approximately 45 trillion multimodal tokens, Inkling introduces controllable reasoning effort, allowing users to trade off latency and reasoning depth depending on the task. ... -
13
Gemopus
Stable fine-tuned Gemma model for structured, clear responses
...While not intended as a production-ready system, it serves as a high-quality local assistant for structured writing, summarization, and coding tasks. Limitations include potential hallucinations in complex domains and weaker performance compared to larger frontier models. -
14
Hermes 4
Hermes 4 FP8: hybrid reasoning Llama-3.1-405B model by Nous Research
Hermes 4 405B FP8 is a cutting-edge large language model developed by Nous Research, built on Llama-3.1-405B and optimized for frontier reasoning and alignment. It introduces a hybrid reasoning mode with explicit <think> segments, enabling the model to deliberate deeply when needed and switch to faster responses when desired. Post-training improvements include a vastly expanded corpus with ~60B tokens, boosting performance across math, code, STEM, logic, creativity, and structured outputs. ... -
15
GLM-5.3-Flash
Efficient 320B multimodal MoE model for coding and autonomous agents
...It also uses Manifold-Constrained Hyper-Connections (mHC) to improve scaling efficiency and was pretrained on a 30-trillion-token multimodal corpus. GLM-5.3-Flash supports text and image inputs and is particularly optimized for coding and autonomous agent workloads, approaching larger frontier models on related benchmarks while improving over GLM-5.2. It supports local deployment through SGLang, vLLM, TokenSpeed, and KTransformers and is released under the MIT license. -
16
Gemma 4
Google’s flagship dense multimodal model for coding and reasoning
...Gemma 4 31B supports native function calling, structured outputs, and more than 140 languages, making it suitable for enterprise assistants, coding agents, document analysis, and multilingual applications. Google positions it as a frontier-level model that can run on consumer GPUs and workstations while achieving leading results across reasoning, mathematics, coding, and multimodal benchmarks. -
17
DeepSeek-V4-Pro
Flagship MoE model for advanced reasoning, coding, and agents
DeepSeek-V4-Pro is a flagship open-weight Mixture-of-Experts language model designed for high-performance reasoning, coding, and agent-based workflows at scale. It features approximately 1.6 trillion total parameters with around 49B activated during inference, enabling strong efficiency while maintaining frontier-level capability. The model supports an ultra-long context window of up to 1 million tokens, making it highly suitable for long-document reasoning, large codebases, and complex multi-step tasks. Architecturally, it introduces optimizations to reduce compute and memory costs while improving stability across long sequences. DeepSeek-V4-Pro is positioned as the high-end variant of the V4 family, outperforming most open-source models in areas such as agentic coding, STEM reasoning, and world knowledge, and approaching the performance of leading closed-source systems. ... -
18
Laguna M.1
Flagship Poolside model for agentic coding and software engineering
...Trained from scratch on roughly 30 trillion tokens using Poolside’s in-house “Model Factory” pipeline, the model focuses on complex software development tasks, repository-scale reasoning, tool use, and multi-step agent execution. Laguna M.1 was designed to compete with leading frontier coding models on benchmarks such as SWE-Bench, Terminal-Bench, and other agentic engineering evaluations. It supports reasoning, tool calling, and long-context workflows, making it suitable for autonomous coding agents, software maintenance, debugging, and large-scale development projects. -
19
Mistral Large 3 675B Instruct 2512 Eagle
Speculative-decoding accelerator for the 675B Mistral Large 3
...It works alongside the primary 675B instruct model, enabling faster response times by predicting several tokens ahead using Mistral’s Eagle speculative method. Built on the same frontier-scale multimodal Mixture-of-Experts architecture, it complements a system featuring 41B active parameters and a 2.5B-parameter vision encoder. The Eagle variant is specialized rather than standalone, serving as a performance accelerator for production-grade assistants, agentic workflows, long-context applications, and retrieval-augmented reasoning pipelines. ... -
20
Mistral Large 3 675B Instruct 2512 NVFP4
Quantized 675B multimodal instruct model optimized for NVFP4
Mistral Large 3 675B Instruct 2512 NVFP4 is a frontier-scale multimodal Mixture-of-Experts model featuring 675B total parameters and 41B active parameters, trained from scratch on 3,000 H200 GPUs. This NVFP4 checkpoint is a post-training-activation quantized version of the original instruct model, created through a collaboration between Mistral AI, vLLM, and Red Hat using llm-compressor. -
21
Ministral 3 14B Reasoning 2512
High-precision 14B multimodal model built for advanced reasoning tasks
Ministral 3 14B Reasoning 2512 is the largest model in the Ministral 3 series, delivering frontier-level performance with capabilities comparable to the Mistral Small 3.2 24B model. It pairs a 13.5B-parameter language model with a 0.4B vision encoder, enabling strong multimodal reasoning across both text and images. This version is specifically post-trained for reasoning tasks, making it highly effective for math, coding, STEM workloads, and complex multi-step problem-solving. ... -
22
Ministral 3 14B Instruct 2512
Efficient 14B multimodal instruct model with edge deployment and FP8
Ministral 3 14B Instruct 2512 is the largest model in the Ministral 3 family, delivering frontier performance comparable to much larger systems while remaining optimized for edge-level deployment. It combines a 13.5B-parameter language model with a 0.4B-parameter vision encoder, enabling strong multimodal understanding in both text and image tasks. This FP8 instruct-tuned variant is designed specifically for chat, instruction following, and agentic workflows with robust system-prompt adherence. ... -
23
Mistral Large 3 675B Instruct 2512
Frontier-scale 675B multimodal instruct MoE model for enterprise AIMis
Mistral Large 3 675B Instruct 2512 is a state-of-the-art multimodal granular Mixture-of-Experts model featuring 675B total parameters and 41B active parameters, trained from scratch on 3,000 H200 GPUs. As the instruct-tuned FP8 variant, it is optimized for reliable instruction following, agentic workflows, production-grade assistants, and long-context enterprise tasks. It incorporates a massive 673B-parameter language MoE backbone and a 2.5B-parameter vision encoder, enabling rich multimodal...
- Previous
- You're on page 1
- Next
