United States
KI-Organisationen und -Modelle aus United States
Aktive Organisationen
48
Erfasste Modelle
873
Neueste Releases aus United States
Oktober 2026
Google's strongest Gemini 4 model, built on Gemini 4 Argon for long-horizon agentic engineering and cybersecurity defense.
Drex Max 1.1 is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities. Nace says it generalises better out of distribution than Drex Max. Supports text and image inputs.
Liquid AI open-weight 1.92B decision model built on LFM3-VL-2B. Answers classification, routing and scoring questions over text, images and video in one forward pass, without generating tokens.
Liquid AI open-weight 2.17B decision model built on LFM3-Omni-2B. Adds audio input to d1-2B for classification, routing and scoring in one forward pass, without generating tokens.
Liquid AI open-weight 1.73B text model with a 256K-token context window, built for local agents, tool calling and structured output. Reasons before answering by default.
Liquid AI open-weight 2B audio model for speech recognition, text-to-speech, speech translation and spoken tool calling in 18 languages, answering in text, speech or both.
Liquid AI's first omni model: a 2.17B open-weight model that understands text, images, video and audio and answers in text, speech or interleaved text and speech.
Liquid AI open-weight 1.92B vision-language model that reads text, images and video at native resolution, with a 128K-token context window and optional reasoning.
A fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.
Organisationen aus United States
193 Modelle
185 Modelle
50 Modelle
48 Modelle
44 Modelle
44 Modelle
36 Modelle
34 Modelle
34 Modelle
24 Modelle
19 Modelle
16 Modelle
14 Modelle
12 Modelle
10 Modelle
9 Modelle
8 Modelle
7 Modelle
7 Modelle
7 Modelle
6 Modelle
6 Modelle
6 Modelle
5 Modelle
5 Modelle
5 Modelle
4 Modelle
4 Modelle
4 Modelle
3 Modelle
3 Modelle
3 Modelle
3 Modelle
2 Modelle
2 Modelle
2 Modelle
2 Modelle
2 Modelle
1 Modell
1 Modell
1 Modell
1 Modell
1 Modell
0 Modelle
0 Modelle
0 Modelle
0 Modelle
0 Modelle
Modelle aus United States
Durchsuche den vollständigen, nach Organisationen gruppierten Katalog.
873 Modelle
Abliteration AI
2 ModelleAbliteration AI's abliterated multimodal reasoning model based on DeepSeek V4.1 Flash 1017. It accepts text and images and supports tool calling and structured outputs, with a one-million-token context window through Venice.
Abliteration AI’s proprietary, text-only reasoning model derived from GLM 5.3, with a one-million-token context window.
AllenAI
9 ModelleMolmo 2 4B is a multimodal model from AllenAI that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
Molmo 2 8B is a multimodal model from AllenAI that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
Bolmo 1B is a language model from AllenAI for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Bolmo 7B is a language model from AllenAI for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Olmo 3.1 32B Instruct is a language model from AllenAI for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Olmo 3.1 32B Think is a large-scale, 32-billion-parameter model purpose-built for deep reasoning, complex logic chains and advanced instruction-following scenarios.
Amazon
19 ModelleNova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text. Nova 2 Lite demonstrates standout capabilities in processing...
Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text. It supports turning still images into animated clips as well as prompt-driven video workflows.
Nova 2 Omni is a speech-to-text model from Amazon for transcription and spoken-language understanding workflows. It is suited to captioning, meeting notes, transcription, and speech analytics pipelines.
Nova 2 Pro is a speech-to-text model from Amazon for transcription and spoken-language understanding workflows. It is suited to captioning, meeting notes, transcription, and speech analytics pipelines.
Nova 2 Sonic is a speech-to-text model from Amazon for transcription and spoken-language understanding workflows. It is suited to captioning, meeting notes, transcription, and speech analytics pipelines.
Nova Multimodal Embeddings is an embedding model from Amazon for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Anthropic
36 ModelleClaude Haiku 5.5 is Anthropic's fastest current model, released on 2026-10-07, for high-volume, latency-sensitive tasks such as classification, extraction, and routing. It supports adaptive thinking, a 1M-token context window, and 128K-token output.
Claude Sonnet 5.5 is Anthropic's Sonnet-class model balancing speed and intelligence, released on 2026-09-28. It combines a 1M-token context window and 128K-token output limit with adaptive thinking.
Claude Opus 5.5 is Anthropic's model for long-running agentic coding and knowledge work, released on 2026-09-22. It always runs with adaptive thinking (default effort medium), and has a 1M-token context window and 128K-token output limit.
Claude Mythos 5.1 is Anthropic's restricted-access Mythos release for approved Project Glasswing customers. Phaseo tracks the model but does not expose a public gateway route.
Claude Fable 5.1 is Anthropic's model for demanding reasoning and long-horizon agentic work, and the successor to [Claude Fable 5](/anthropic/claude-fable-5). It always runs with adaptive thinking (default effort high), and has a 1M-token context window and 128K-token output limit.
Claude Opus 5 is an Anthropic Opus-class model for complex reasoning, long-horizon agentic coding, and high-autonomy work, released on 2026-07-24. It has a 1M token context window and 128K max output. [Claude Opus 5.5](/anthropic/claude-opus-5.5) is its successor.
Arcee AI
5 ModelleTrinity Large Thinking is a powerful open source reasoning model from the team at Arcee AI. It shows strong performance in PinchBench, agentic workloads, and reasoning tasks.
Trinity Large is a frontier-scale open-weight language model from Arcee, built as a 400B-parameter sparse Mixture-of-Experts with 13B active parameters per token using 4-of-256 expert routing.
Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token. It is a smaller-footprint option suited to tighter latency or deployment budgets.
Trinity Nano Preview is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token. This entry tracks a preview release and capabilities may change before general availability.
Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA. Unlike many 70 B peers, it retains the 128 k...
Base Labs
3 ModelleA 2B decision model from Base Labs, fine-tuned from Qwen3.5-2B for fast, low-cost classification, routing, and scoring. Returns probabilities over predefined answers without generating text.
A 9B decision model from Base Labs, fine-tuned from Qwen3.5-9B for classification, routing, and scoring. Returns probabilities over predefined answers without generating text.
A 27B decision model from Base Labs, fine-tuned from Qwen3.5-27B for classification, routing, and scoring. Returns probabilities over predefined answers without generating text.
Cloudflare
4 ModelleTyped extraction with confidence scores and missing-value probabilities.
Cloudflare 30B-A3B multimodal model for typed decisions over text, images, audio, and video.
Cloudflare 27B multimodal model for typed Noul, Choice, and Score decisions.
Cloudflare 9B multimodal model for typed Noul, Choice, and Score decisions.
Cogito
1 ModellCogito 671b V2.1 is a language model from Cogito for chat, writing, analysis, and general-purpose automation.
CrofAI
7 ModelleGreg 2 Super is CrofAI's high-capability Greg 2 text model for chat, writing, analysis, and general-purpose text generation.
Greg 2 Ultra is CrofAI's premium Greg 2 text model for higher-capability responses across long-context chat, writing, and analysis tasks.
Greg RP is CrofAI's Greg roleplay variant, designed for character-driven text generation and conversational storytelling.
Greg 1 is CrofAI's general-purpose language model for chat, writing, analysis, and everyday automation.
Greg 1 Mini is a lightweight CrofAI language model for fast chat, writing, and general-purpose text generation.
Greg 1 Super is CrofAI's premium Greg 1 variant, tuned for higher-capability responses while keeping the same long text context window.
Cursor
3 ModelleComposer 2.5 is a language model from Cursor optimized for long-horizon coding and agentic software tasks. It accepts text inputs and produces text outputs.
Composer 1.5 is a language model from Cursor for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Composer 1 is a language model from Cursor for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Essential AI
1 ModellRnj-1 is a language model from Essential AI for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Fastino
2 ModelleFastino's finance-specialized Nemotron 3.5 Lightning model for inference and LoRA training on Pioneer.
Fastino's healthcare-specialized Nemotron 3.5 Lightning model for inference and LoRA training on Pioneer.
Fireworks
1 ModellEmber-1 is Fireworks' specialized model built on Kimi K3. It produces shorter reasoning traces, using approximately 40% fewer tokens while maintaining comparable quality across Fireworks' evaluations.
Fish Audio
7 ModelleFish Audio's production text-to-speech model with multi-speaker generation, inline expression controls, and support for 83 languages.
Fish Audio's production text-to-speech model with multi-speaker generation, inline expression controls, and support for 83 languages.
Fish Audio's open-weight multilingual text-to-speech model, supporting more than 80 languages and low-latency streaming generation.
The compact 0.5B-parameter open-weight edition of Fish Audio's S1 expressive text-to-speech model.
Fish Audio's 4B-parameter expressive text-to-speech model with 13-language support and more than 64 expression controls.
Fish Audio's speech-to-text model for transcription with timestamped output.
Google's strongest Gemini 4 model, built on Gemini 4 Argon for long-horizon agentic engineering and cybersecurity defense.
Gemini 4 Argon with Contributor pricing listed by the Gemini API.
Google frontier reasoning model for software engineering, enterprise knowledge work, multimodal understanding, and cybersecurity defense.
EmbeddingGemma 2 is Google DeepMind's open multimodal embedding model for text, code, images, video, and audio retrieval. Built on Gemma 4, it maps interleaved inputs into a shared 768-dimensional space with an 8,192-token context. Its modular encoders scale from 270M parameters for text to 740M for all modalities; Matryoshka embeddings support 512, 256, and 128 dimensions. Released under Apache 2.0 for on-device search, RAG, classification, and clustering.
High-quality image generation and editing model with enhanced factuality, visual fidelity, and multi-turn consistency at Flash scale.
A text-to-speech audio model optimized for high-throughput, low-latency, and cost-efficient conversational speech.
Hexgrad
1 ModellKokoro-82M is an open-weight multilingual, multi-voice text-to-speech model optimized for fast speech synthesis.
IBM
50 ModelleGranite 4.2 30B is IBM's flagship reasoning language model with support for thinking modes, tool calling, and text generation.
Granite 4.2 3B is a compact reasoning language model from IBM with support for thinking modes, tool calling, and text generation.
Granite 4.2 8B is a reasoning language model from IBM with support for thinking modes, tool calling, and text generation.
Granite 4.1 30B is a language model from IBM for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Granite 4.1 3B is a language model from IBM for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Granite 4.1 8B is a language model from IBM for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Inception
6 ModelleMercury 3 is Inception's flagship diffusion language model for fast reasoning, chat, coding, and agent workflows. It supports a 1M-token context window, up to 131,072 output tokens, tool calling, and structured outputs.
Mercury 2.5 is Inception's reasoning diffusion language model for fast, high-quality text generation, coding, and agent workflows. It accepts text inputs and produces text outputs.
Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM). It accepts text inputs and produces text outputs.
Mercury Edit 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM). It accepts text inputs and produces text outputs.
Mercury Decide is a structured decision model from Inception for classification, routing and scoring. It evaluates named yes/no, choice or score questions against a shared text or JSON state and returns typed answers with probabilities instead of generated text.
Mercury Decide (Free) is a structured decision model from Inception for classification, routing and scoring. It evaluates named yes/no, choice or score questions against a shared text or JSON state and returns typed answers with probabilities instead of generated text.
Inference.net
2 ModelleSchematron V2 Small is Inference.net's schema-focused serverless model for complex structured extraction and long pages.
Schematron V2 Turbo is Inference.net's lower-cost, higher-throughput schema-focused serverless model.
Inflection AI
3 ModelleInflection 3 Pi powers Inflection's [Pi](https://pi.ai) chatbot, including backstory, emotional intelligence, productivity, and safety. It has access to recent news, and excels in scenarios like customer support and roleplay.
ai) chatbot, including backstory, emotional intelligence, productivity, and safety. It has access to recent news, and excels in scenarios like customer support and roleplay. This entry tracks a preview release and capabilities may change before general availability.
Productivity Pi is optimized for following instructions. It is better for tasks requiring JSON output or precise adherence to provided guidelines. It has access to recent news.
Liquid AI
24 ModelleLiquid AI open-weight 1.92B decision model built on LFM3-VL-2B. Answers classification, routing and scoring questions over text, images and video in one forward pass, without generating tokens.
Liquid AI open-weight 2.17B decision model built on LFM3-Omni-2B. Adds audio input to d1-2B for classification, routing and scoring in one forward pass, without generating tokens.
Liquid AI open-weight 1.73B text model with a 256K-token context window, built for local agents, tool calling and structured output. Reasons before answering by default.
Liquid AI open-weight 2B audio model for speech recognition, text-to-speech, speech translation and spoken tool calling in 18 languages, answering in text, speech or both.
Liquid AI's first omni model: a 2.17B open-weight model that understands text, images, video and audio and answers in text, speech or interleaved text and speech.
Liquid AI open-weight 1.92B vision-language model that reads text, images and video at native resolution, with a 128K-token context window and optional reasoning.
Meta
44 ModelleLimited-time free Together offer for Muse Spark Open. Ends October 18, 2026 at 19:00 UTC. Requires an eligible Together account.
Muse Spark Open is Meta’s open-weight multimodal reasoning model for coding and long-horizon agentic work. It uses 816B total parameters with 42B active per token, supports a 1M-token context window, and includes additional safety post-training for standalone deployment.
Meta’s higher-capability Muse Wonder variant for the hardest coding and long-running agentic work, running at maximum reasoning. Available only through Muse Code with Power Usage or Maximum Usage subscriptions.
Muse Wonder is Meta's agentic model for coding and multi-step work, with native tool calling and MCP support. It plans long tasks, delegates work, and recovers from errors.
Muse Spark 1.3 is Meta's agentic model for long-horizon coding and multi-step work. It improves tool, browser, and computer use across agent harnesses, adds native tool calling and MCP support, and uses fewer tokens than Muse Spark 1.2.
Muse Spark 1.3 Contributor has the same model and capabilities as Muse Spark 1.3 at lower prices. Inputs and outputs may be used to train and improve Meta's AI models.
Microsoft
48 ModelleDecision-scoring model for routing, classification, prioritization, verification, and workflow control. Scores fixed answer options with calibrated probabilities.
MAI-Transcribe-2-Streaming is a low-latency speech-to-text model from Microsoft for real-time transcription. Audio is sent as a continuous stream and incremental transcripts are returned while the speaker talks, with intermediate results updated and final results confirming each segment.
MAI-Voice-2.1 is a text-to-speech model from Microsoft for expressive, high-fidelity speech generation. It supports 23 languages, SSML emotion and style control, gated instant voice cloning, and long-form generation that keeps speaker consistency across audiobooks, podcasts and narration.
MAI-Voice-2.1-Flash is a text-to-speech model from Microsoft for low-latency, real-time speech generation. It produces expressive speech across 23 languages with emotion and style control through SSML, and supports gated instant voice cloning.
MAI-Image-2.6 is a diffusion-based image generation and editing model from Microsoft for text-to-image generation and image-to-image edits. Prompts can be up to 32,000 tokens, edits accept up to five reference images, and output is PNG up to 2,359,296 total pixels.
MAI-Image-2.6-Flash is a diffusion-based image generation and editing model from Microsoft with the same text-to-image and image-to-image capabilities as MAI-Image-2.6, offered as a faster variant. Prompts can be up to 32,000 tokens, and edits accept up to five reference images.
Morph
6 ModelleMorph Compactor is a language model from Morph for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Morph Embedding V4 is an embedding model from Morph for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Morph Rerank V4 is a reranking model from Morph for improving retrieval quality by rescoring candidate results against a query. Use it to rescore retrieved passages, products, or documents before final ranking or generation.
Morph's fastest apply model for code edits. ~10,500 tokens/sec with 96% accuracy for rapid code transformations. It is tuned for lower latency and cost-sensitive production use.
Morph's high-accuracy apply model for complex code edits. ~4,500 tokens/sec with 98% accuracy for precise code transformations. It targets higher-capability workloads where depth, quality, or scale matter more than raw speed.
Morph WarpGrep V2.1 is a coding model from Morph for code generation, edits, and software tasks. It accepts text inputs and produces text outputs.
Nace.AI
6 ModelleDrex Max 1.1 is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities. Nace says it generalises better out of distribution than Drex Max. Supports text and image inputs.
Drex Max is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities. Supports text and image inputs.
Drex 2.0 is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities.
Drex 1.5 is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities.
Drex 1.0 is a Nace decision model that answers typed choice, yes/no and score questions with calibrated probabilities.
Retired Nace decision model; upstream requests are served by Drex 1.5.
Nous
16 ModelleNousCoder 14B is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Nomos 1 is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Hermes 4.3 36B is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Hermes 4 14B is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Hermes 4 405B is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Hermes 4 70B is a language model from Nous for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Nvidia
34 ModelleNVIDIA’s open 120B-parameter vision-language MoE model with 12B active parameters, built for reasoning, coding, tool use, document understanding and video understanding.
Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.
Nemotron 3.5 Lightning is NVIDIA's open 30B-parameter mixture-of-experts reasoning model with 3B active parameters, a hybrid Mamba-Transformer architecture, and support for high-throughput agentic workloads.
Nemotron 3 Embed 1B BF16 is a multilingual text embedding model for retrieval, semantic search, and RAG workflows.
Nemotron 3 Embed 8B BF16 is a multilingual text embedding model for retrieval, semantic search, and RAG workflows.
NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.
OpenAI
193 ModelleA fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.
GPT Live 1.1 is an update to OpenAI's full-duplex GPT Live voice model for ChatGPT Voice, Codex and dots, adding 13 new voices across six non-English languages, better accents, more natural filler while it works and better instruction following. API access is coming soon.
GPT-6.1 Sol with OpenAI Pro reasoning mode for difficult tasks.
Near-Astra performance for complex coding, computer use, and professional work at a lower cost.
GPT-6 Luna Pro is the same underlying GPT-6 Luna model served with reasoning.mode set to pro for higher-quality responses on complex tasks.
GPT-6 Sol Pro is the same underlying GPT-6 Sol model served with reasoning.mode set to pro for higher-quality responses on complex tasks.
Perplexity
14 ModelleExclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system. It is designed for deeper reasoning and analysis. Pricing is based...
Premier reasoning offering powered by DeepSeek R1 with Chain of Thought (CoT).
Premier search offering with search grounding, supporting advanced queries and follow-ups.
Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics. It autonomously searches, reads, and evaluates sources, refining its approach as it gathers...
Lightweight offering with search grounding, quicker and cheaper than Sonar Pro.
Perplexity's most thorough research agent. It runs many rounds of web search and page reading on a frontier model, then writes a long, cited report. Requests can take several minutes.
Phala
2 ModelleGemma-4 26B-A4B Uncensored (Heretic) is available through Phala's confidential GPU TEE inference catalog.
Qwen3.6 35B-A3B Uncensored (Aggressive) is available through Phala's confidential GPU TEE inference catalog.
Poolside
8 ModelleLaguna S 2.1 is Poolside's 118B total parameter, 8B active parameter Mixture-of-Experts model for long-horizon agentic coding and reasoning workloads.
Laguna S 2.1 is Poolside's 118B total parameter, 8B active parameter Mixture-of-Experts model for long-horizon agentic coding and reasoning workloads.
Laguna XS 2.1 is Poolside's upgraded 33B total parameter, 3B active parameter Mixture-of-Experts coding model for agentic software engineering and long-horizon local workflows.
Laguna XS 2.1 is Poolside's upgraded 33B total parameter, 3B active parameter Mixture-of-Experts coding model for agentic software engineering and long-horizon local workflows.
Laguna XS.2 is the second-generation model in the XS size class from [Poolside](https://poolside.ai), their efficient coding agent series. It accepts text inputs and produces text outputs.
Laguna XS.2 is the second-generation model in the XS size class from [Poolside](https://poolside.ai), their efficient coding agent series. It accepts text inputs and produces text outputs.
Prime Intellect
2 ModelleIntellect 3.1 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL).
Intellect 3 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL).
Reka AI
5 ModelleReka Flash 3 is a general-purpose, instruction-tuned large language model with 21 billion parameters, developed by Reka. It excels at general chat, coding tasks, instruction-following, and function calling. Featuring a...
Reka's compact multimodal model, also exposed by the versioned reka-edge-2603 alias.
The versioned Reka Edge 2603 model exposed by Reka's public Chat API.
Reka's fast multimodal model for text, image, short-video, audio, and PDF understanding.
Reka's web research model for grounded answers with reasoning steps, citations, streaming, and structured output.
Relace
4 ModelleRelace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files. It can apply updates from GPT-4o, Claude, and others into your files at...
Relace Compact compresses long coding-agent traces while preserving the message format and important details. It accepts OpenAI Chat Completions, OpenAI Responses, or Anthropic Messages traces and returns the same format.
Relace Rank is a code reranker that scores each file's relevance to a natural-language coding request and returns files ordered by relevance.
The relace-search model uses 4-12 `view_file` and `grep` tools in parallel to explore a codebase and return relevant files to the user request. It is aimed at retrieval-heavy assistants, browsing, and research synthesis.
Runway
12 ModelleRunway Aleph 2.0 is an in-context video editing model from Runway. It applies text instructions and keyframe-guided edits across existing footage while preserving details that are not meant to change....
Runway Gen-4.5 generates video from text or a single first-frame image, with durations from 2 to 10 seconds.
Runway Act-Two transfers a driving performance to a character image or video.
Runway Gen-3 Alpha Turbo was a 720p video generation model retired from the API on July 30, 2026.
Runway Gen-4 Aleph was a video editing model retired from the API on July 30, 2026.
Runway Gen-4 Image generates 720p or 1080p images from text and reference images.
SpaceXAI
44 ModelleGrok Voice 3.0 is a realtime speech-to-speech model for voice assistants and customer support agents, using the xAI Voice Agent WebSocket API.
SpaceXAI's frontier model for coding, agentic tasks, and knowledge work. Accepts text and image input, produces text, and has a 500,000-token context window.
Grok Voice Transcribe 2.0 is xAI's next-generation speech-to-text model for batch transcription through the native /v1/stt API.
Grok 4.6 is SpaceXAI's model for coding, agentic tasks, and knowledge work. It supports text and image input, a 500k-token context window, configurable reasoning, function calling, structured outputs, and xAI's server-side web search, X search, and code execution tools.
Grok Imagine Image 2.0 is SpaceXAI's image generation and editing model. It accepts text and image inputs, produces 1K or 2K images at low, medium, or auto quality, supports batch generation and up to five source images for editing, and is available through the xAI Imagine API.
Grok Voice Think Fast 2.0 is xAI's next-generation realtime speech-to-speech model, with improved reasoning, transcription accuracy, conversational ability and tool-use reliability.
Suno
10 ModelleSuno V6 Wild is Suno's exploratory v6 music model. Suno describes it as less predictable and more varied, producing unexpected, textured, and ambitious results, and makes it available to Pro and Premier subscribers.
Suno V6 is Suno's flagship v6 music-generation model. It is described as reliable and precise, with polished music generation across genres and styles, and is available to Pro and Premier subscribers.
Suno V6 Mini is the faster, more efficient member of Suno's v6 music-model generation. Suno says it is available to everyone and makes it easier to turn an idea into a song.
Suno v5.5 is a proprietary music-generation model with improved expressiveness and support for personalized Voices and Custom Models in Suno's product.
Suno v4.5-all is a proprietary music-generation model released as Suno's free model, with richer, fuller, and more dynamic output.
Suno v5 is a proprietary music-generation model with clearer audio, more natural vocals, improved creative control, and stronger genre understanding.
Thinking Machines Lab
3 ModelleInkling-Small is Thinking Machines Lab's efficient open-weights multimodal mixture-of-experts model with 276B total parameters and 12B active parameters, designed for faster inference and lower cost.
Inkling 256K is the extended context variant of Inkling, a large MoE hybrid reasoning model from Thinking Machines with audio and vision input support and a 256K context window.
Inkling is Thinking Machines Lab's open-weights multimodal mixture-of-experts model for customizable AI systems. It accepts text, image, and audio inputs and produces text outputs.
Together
1 ModellTogether's experimental Qwen3.5-4B decision model. Supply state, question, and 2-24 labeled options; request non-thinking generation and map the returned A-X label to its semantic key.
Venice
7 ModelleGemma 4 26B A4B Uncensored running in a Trusted Execution Environment (TEE). An uncensored variant of Google's Gemma 4 MoE model with 25.2B total / 3.8B active parameters, supporting multimodal input across text and images, with hardware attestation evidence available for independent verification.
Qwen3.6 35B A3B Uncensored is an uncensored Qwen3.6 MoE variant offered by Venice in a Trusted Execution Environment, with 128K context and text input on this route.
Gemma 4 Uncensored is an uncensored variant of Google Gemma 4 26B, a Mixture-of-Experts model with 26B total parameters and only 4B active per token. Fine-tuned for uncensored chat without content filtering, it supports 256K context, coding, and general-purpose conversation.
Venice Uncensored 1.2 is designed for maximum creative freedom and authentic interaction. Built for open-ended exploration, roleplay, and unfiltered dialogue with improved capabilities over 1.1.
Optimized for creative roleplay scenarios with maximum freedom. Designed for immersive storytelling, character interactions, and open-ended creative writing.
Venice Uncensored is a language model from Venice for chat, writing, analysis, and general-purpose automation. It accepts text inputs and produces text outputs.
Vercel
4 Modellev0 1.0 md is a multimodal model from Vercel that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
v0 1.5 lg is a multimodal model from Vercel that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
v0 1.5 md is a multimodal model from Vercel that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
v0 1.5 sm is a multimodal model from Vercel that can work across text and visual inputs. It accepts text and image inputs and produces text outputs.
Voyage
34 ModelleRerank 3 is a reranking model from Voyage AI that scores and reorders candidate documents by relevance to a query. It has a 32K-token context length and supports instruction-following.
Rerank 3 Lite is a reranking model from Voyage AI that scores and reorders candidate documents by relevance to a query, optimized for latency-sensitive applications. It has a 32K-token context length and supports instruction-following.
Voyage 4 is an embedding model from Voyage for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Voyage 4 Large is an embedding model from Voyage for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Voyage 4 Lite is an embedding model from Voyage for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Voyage Multimodal 3.5 is an embedding model from Voyage for vector search and semantic retrieval workflows. It is aimed at semantic search, retrieval, clustering, and RAG pipelines.
Windsurf
5 ModelleSWE-1 is Windsurf's first in-house agentic coding model.
SWE-1 Mini powers passive code suggestions in Windsurf Tab and is optimized for real-time latency.
SWE-1.5 is Windsurf's previous frontier agentic coding model for software engineering tasks.
SWE-1.6 is Windsurf's in-house frontier model for software engineering agents, optimized for agentic coding tasks and model UX.
SWE-grep powers context retrieval and Fast Context in Windsurf.