Map of artificial intelligence

Discover systems, capabilities and who builds them.

An editorial catalog for exploring AI by technology and problem, with verified profiles and clear paths.

Verified models

Profiles with version, access, modalities, published pricing and verification date.

OPENAI

GPT‑6 Astra

Generalist frontier model oriented towards complex tasks, use of tools and highly demanding professional work.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Opus 5

Model oriented to code and knowledge work, with controls that allow effort to be adjusted to each task.

Text and reasoning · Paid proprietary GOOGLE

Gemini 3.8 Flash

Flash model for long-running software engineering, autonomous agents, and complex business flows.

Text and reasoning · Paid proprietary OPENAI

GPT‑5.6 Sol

Complex work, coding, and tool-using agents.

Text and reasoning · Paid proprietary OPENAI

GPT‑5.6 Terra

Tasks requiring a balance of quality, speed, and cost.

Text and reasoning · Paid proprietary OPENAI

GPT‑5.6 Luna

High volume, subagents, and structured extraction.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Fable 5.1

Long-running reasoning and agents.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Sonnet 5

Coding, research, and professional tool-based workflows.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Haiku 4.5

High volume, subagents, and structured extraction.

Text and reasoning · Paid proprietary GOOGLE

Gemini 3.1 Pro

Multimodal reasoning, coding, and complex problems.

Text and reasoning · Paid proprietary XAI

Grok 4.6

Coding, research, and professional tool-based workflows.

Text and reasoning · Paid proprietary DEEPSEEK

DeepSeek V4.1 Flash

Agents, coding, and open deployment at low cost.

Text and reasoning · Paid proprietary MISTRAL AI

Mistral Medium 3.5

Multimodal coding and self-controlled deployment.

Text and reasoning · Paid proprietary META

Llama 4 Maverick

Open multimodality and self-controlled deployment.

Text and reasoning · Paid proprietary COHERE

Command A+

Enterprise RAG, multilingual use, and private deployment.

Text and reasoning · Paid proprietary AMAZON

Amazon Nova 2 Lite

Long documents, video, and enterprise tasks on AWS.

Text and reasoning · Paid proprietary ALIBABA QWEN

Qwen3‑Max Thinking

Multimodal reasoning, coding, and complex problems.

Text and reasoning · Paid proprietary Z.AI

GLM‑5.1

Coding, research, and professional tool-based workflows.

Text and reasoning · Paid proprietary MINIMAX

MiniMax M2.7

Agents, coding, and open deployment at low cost.

Text and reasoning · Paid proprietary OPENAI

GPT‑5.5

Previous OpenAI generation for code and professional work, retained for comparing migrations and cost.

Text and reasoning · Paid proprietary OPENAI

GPT‑4.1

Long-context non-reasoning model, useful as a historical reference for code, instruction following, and predictable latency.

Code · Paid proprietary OPENAI

gpt‑oss‑120b

Reasoning model with open weights for local or private execution; fits on one H100 GPU according to OpenAI.

Text and reasoning · Open weights OPENAI

GPT‑Image‑2.5 Sunburst

OpenAI's main model for generating and editing images with complex instructions and high-quality visual production.

Image · Paid proprietary OPENAI

Sora 2 Pro

Historical video model with synchronized audio, useful for comparing the evolution of the Sora family.

Video · Retired OPENAI

GPT‑Live 1

Real-time voice model for expressive conversations, natural interruptions, and tool-enabled workflows.

Voice and audio · Paid proprietary OPENAI

GPT‑Transcribe

Specialized high-accuracy transcription model for files and real-time audio inputs.

Voice and audio · Paid proprietary OPENAI

text‑embedding‑3‑large

OpenAI's highest-capacity embedding model for semantic search, classification, clustering, and RAG.

Embeddings · Paid proprietary ANTHROPIC

Claude Opus 4.5

Previous Opus version focused on code, agents, and computer use, retained for migration and historical comparison.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Sonnet 4.5

Previous Sonnet generation for code, agents, and interface automation, with official computer use results.

Code · Paid proprietary GOOGLE

Gemini 3.7 Flash

Previous Flash generation for complex code, agents, and multi-stage execution; useful as a comparison point with Gemini 3.8.

Text and reasoning · Paid proprietary GOOGLE

Nano Banana 2

Production image model that combines generation, editing, and visual reasoning with a balance of quality, cost, and latency.

Image · Paid proprietary GOOGLE

Veo 3.1

Cinematic video model with native audio, scene extension, and control through frames and reference images.

Video · Paid proprietary GOOGLE

Lyria 3.5

Music model for full songs with vocals, lyrics, and control over structure, style, duration, and instrumentation.

Music · Paid proprietary GOOGLE

Gemini Embedding 2

Embedding model that places text, images, video, audio, and PDFs in a shared vector space for search and multimodal RAG.

Embeddings · Paid proprietary GOOGLE

Gemini Robotics ER 2

Model specialized in physical understanding, video, spatial reasoning, and tool orchestration for robotic systems.

Robotics · Paid proprietary META

Llama 4 Scout

Multimodal model with 17B active parameters and 16 experts, designed for context up to 10 million tokens.

Multimodal · Open weights META

Segment Anything Model 3

Vision model for detecting, segmenting, and tracking open concepts in images and video through text or visual examples.

Vision and documents · Open weights DEEPSEEK

DeepSeek V3.2

Open model focused on reasoning and agents, with thinking integrated into tool use.

Text and reasoning · API and open weights DEEPSEEK

DeepSeek R1

Open reasoning model and distilled family, relevant for tracking the evolution of verifiable reasoning.

Text and reasoning · API and open weights MISTRAL AI

Mistral Small 4

Hybrid model that brings together instruction following, reasoning, and code in an efficient option with open weights.

Text and reasoning · API and open weights MISTRAL AI

Voxtral TTS

Multilingual text-to-speech model with streaming and zero-shot cloning from a voice reference.

Voice and audio · API and open weights MISTRAL AI

Mistral OCR 4.1

Specialized OCR service with paragraph bounding boxes, structured block labels, and confidence scores.

Vision and documents · Paid proprietary ALIBABA QWEN

Qwen3 235B A22B

Open MoE model with 235B total parameters and 22B active parameters, with thinking and direct-response modes.

Text and reasoning · API and open weights ALIBABA QWEN

Qwen3‑Coder 480B A35B

Open MoE model specialized in software engineering, repository navigation, and extended tool use.

Code · API and open weights MICROSOFT

Phi‑4 Multimodal

Compact Microsoft model that combines language, vision, and audio, including speech transcription and translation.

Multimodal · Open weights COHERE

Cohere Embed 4

Multimodal embedding model for text, images, and mixed documents with configurable dimensions.

Embeddings · Paid proprietary COHERE

Command R 08‑2024

Previous commercial model optimized for RAG, citations, tools, and high performance on enterprise workloads.

Text and reasoning · Paid proprietary STABILITY AI

Stable Diffusion 3.5 Large

8B-parameter image base model for generation and editing, available via API and for self-deployment under a Stability license.

Image · API and open weights STABILITY AI

Stable Audio 3.0

Commercial model for generating songs, instruments, and effects from text or audio, with inpainting and style transformation.

Music · Paid proprietary ELEVENLABS

Eleven v3

Expressive text-to-speech model with emotional tags, multi-speaker dialogue, and support for more than 70 languages.

Voice and audio · Paid proprietary ELEVENLABS

Scribe v2

Speech recognition model for more than 90 languages, with diarization, word-level timestamps, entities, and audio labels.

Voice and audio · Paid proprietary NVIDIA

NVIDIA Nemotron 3 Ultra

Open 550B-parameter MoE model for orchestration, research, code, and long-duration agents.

Text and reasoning · API and open weights NVIDIA

Nemotron Labs Dory 1.0 4B

NVIDIA text model released on October 16, 2026, with 4 billion parameters and a recurrent architecture. It reuses a block of layers to adjust compute per token without switching checkpoints. English and a validated 256K-token context are documented. It does not generate images, audio or video.

Text and reasoning · Open weights MISTRAL AI

Timestral 5M

Mistral numerical model family released on October 15, 2026 in four sizes: 5M, 28M, 123M and 209M. It performs zero-shot univariate and multivariate probabilistic forecasting. It accepts history, past covariates and known future covariates; it neither chats nor generates text.

Time series · Restricted weights MISTRAL AI

Timestral 28M

Mistral numerical model family released on October 15, 2026 in four sizes: 5M, 28M, 123M and 209M. It performs zero-shot univariate and multivariate probabilistic forecasting. It accepts history, past covariates and known future covariates; it neither chats nor generates text.

Time series · Restricted weights MISTRAL AI

Timestral 123M

Mistral numerical model family released on October 15, 2026 in four sizes: 5M, 28M, 123M and 209M. It performs zero-shot univariate and multivariate probabilistic forecasting. It accepts history, past covariates and known future covariates; it neither chats nor generates text.

Time series · Restricted weights MISTRAL AI

Timestral 209M

Mistral numerical model family released on October 15, 2026 in four sizes: 5M, 28M, 123M and 209M. It performs zero-shot univariate and multivariate probabilistic forecasting. It accepts history, past covariates and known future covariates; it neither chats nor generates text.

Time series · Restricted weights BLACK FOREST LABS

FLUX 3 Image Klein

A 5.6-billion-parameter image model released on October 15, 2026. It generates and edits images from text and references, with bounding-box composition control. Distilled and Base are different checkpoints; neither should be confused with the full FLUX 3 Image model.

Image · Restricted weights BLACK FOREST LABS

FLUX 3 Image Klein Base

A 5.6-billion-parameter image model released on October 15, 2026. It generates and edits images from text and references, with bounding-box composition control. Distilled and Base are different checkpoints; neither should be confused with the full FLUX 3 Image model.

Image · Restricted weights META

Muse Spark Open

Meta multimodal model released on October 15, 2026, with 816 billion total and 42 billion active parameters. It accepts text, images, audio and video and returns text. It has a combined one-million-token context and BF16 and NVFP4 variants.

Multimodal · Restricted weights LIQUID AI

d1-ultra

Liquid AI multimodal decision model released on October 15, 2026. It accepts text and images and returns probabilities for predefined choices, binary questions or scales. It uses the d1-ultra identifier in the decisions API; it does not generate free-form text.

Decisions · Paid proprietary AI2

HiRO-ACEv2

Ai2 system announced on October 14, 2026 combining the ACE2S-X-SHiELD atmospheric emulator with the HiROv2 diffusion model. It transforms roughly 100 km fields into 3 km precipitation, ten-metre wind and sea-level pressure data. The public repository includes both checkpoints, configuration and sample data; its creation on October 12 is not confused with the announcement date. It is not a conversational assistant or a conventional image generator.

Time series · Open weights XAI

Grok Voice 3.0

Proprietary voice model available since October 14, 2026 for bidirectional audio and text conversations. Its versioned identifier is grok-voice-3.0; grok-voice-latest is an alias already pointing to this version, not another model. It replaces the deprecated Grok Voice 2.0 without changing the endpoint, session parameters or events. Do not confuse it with Grok Voice Transcribe 2.0 or the separate text-to-speech service.

Voice and audio · Paid proprietary MICROSOFT

UserRepGen-4B

Microsoft checkpoint based on Qwen3.5-4B that turns text conversations into descriptions of the task and observed behaviour. Weights were uploaded to the official repository on October 14, 2026. It generates representations for a separate simulator: it does not itself produce user turns or execute their task. Although it retains the base model’s multimodal architecture, this release supports text only and was evaluated in English.

Text and reasoning · Open weights MICROSOFT

AesCode 8B

Microsoft's multimodal family turns content requirements and optional visual references into editable HTML/CSS documents. The 8B and 32B checkpoints derive from Qwen3-VL-8B-Instruct and Qwen3-VL-32B-Instruct respectively. They produce code, not raster images or video files. The reference provides composition and style, while the prompt remains the source of truth for text, figures and relationships. The verified public release is October 11, 2026; earlier repository creation is not its release date.

Code · Open weights MICROSOFT

AesCode 32B

Microsoft's multimodal family turns content requirements and optional visual references into editable HTML/CSS documents. The 8B and 32B checkpoints derive from Qwen3-VL-8B-Instruct and Qwen3-VL-32B-Instruct respectively. They produce code, not raster images or video files. The reference provides composition and style, while the prompt remains the source of truth for text, figures and relationships. The verified public release is October 11, 2026; earlier repository creation is not its release date.

Code · Open weights ALIBABA QWEN

Qwen Image 2.1 Turbo

Image generation and editing model published on October 9, 2026. The Qwen/Qwen-Image-2.1-Turbo checkpoint uses Qwen Image 2.1's 7-billion-parameter visual component, accelerated to eight denoising steps. Its API identifier is qwen-image-2.1-turbo. Both access paths are recorded in one profile, without confusing Turbo with Pro or duplicating entries by platform.

Image · Restricted weights and paid API MISTRAL

LIDstral Arabic

Language and dialect classifier publicly released by Mistral on Hugging Face on October 9, 2026. It receives Arabic-script text and distinguishes Modern Standard Arabic, dialects and other languages sharing that writing system. It is neither a generative assistant nor a transcription model. It combines fastText one-vs-all, an MLP classifier and per-class isotonic calibration, returning a label and calibrated probabilities over 51 classes.

Text classification · Open weights TENCENT

Youtu-Parsing-Omni

Tencent's 5-billion-parameter multimodal model, publicly released on Hugging Face on October 9, 2026. It analyzes pages, images, charts, diagrams, geometry, audio and video, producing structured text in an OmniSchema JSON envelope. It does not generate images, music or video. A shared encoder processes modalities before the decoder; the requested task determines the schema and extracted elements.

Multimodal · Restricted weights AI2

Bwen 8B

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights AI2

Blama 8B

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights AI2

Bolmo 1B Stage 1

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights AI2

Bolmo 7B Stage 1

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights AI2

Bwen 8B Stage 1

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights AI2

Blama 8B Stage 1

Ai2 checkpoint included in the public release of October 7, 2026, alongside the publication of Bolmo in Nature. It is an autoregressive model that processes text as UTF-8 bytes through local components, dynamic pooling and a global model. The verified modality is text-to-text: potential future extensions to images or audio are not available capabilities. The repository creation date in August is not used as the release date.

Text and reasoning · Open weights GOOGLE

Gemini 4 Argon

Google model generally available since October 13, 2026. It accepts text, images, audio and video and generates text, including code; it does not generate images, audio or video. It uses a MoE architecture and a shared 1,048,576-token window. Robotics applications use this same model, not a separate endpoint.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Haiku 5.5

Anthropic model released on October 7, 2026 for frequent, cost- and latency-sensitive tasks. It accepts text and images and returns text, including code; it does not generate images, audio or video. Its one-million-token window and adaptive thinking expand the Haiku family without making it a universal replacement for Sonnet or Opus on complex work.

Text and reasoning · Paid proprietary GOOGLE

Gemini Nano Banana 2.1

Google model for image generation and conversational editing, generally available since October 6, 2026. It replaces Nano Banana 2 and should not be confused with Nano Banana Pro. It accepts text, images, video and PDF as context; it produces images and text, not video or audio. Google reports improvements in visuals, instruction following, text rendering and consistency across edits.

Image · Paid proprietary IBM

Granite Timeseries Ensemble R1

Time-series forecasting ensemble published by IBM on October 5, 2026, according to the official repository's public publication event. It combines PatchTST-FM-r1, PatchTST-FM-r2, FlowState-r1.1 and TTM-r3. It distributes reproducible recipes, not merged weights or a newly trained checkpoint: the four members are downloaded from their respective repositories at runtime.

Time series · Open ensemble ALEPH ALPHA

Kolibri 1

Aleph Alpha's text and code model, officially released on October 3, 2026, focused on German and English. Its MoE architecture has 78.1 billion total parameters and activates about 3.46 billion per token. It supports configurable reasoning and tool calling. FP8 and BF16 weights are released under Apache 2.0; they are two precisions of the same family, not two independent models.

Text and reasoning · Open weights ALIBABA QWEN

Qwen Image 2.1 Pro

Image generation and editing service added to Alibaba Cloud's international catalog on October 2, 2026. The Qwen Image 2.1 family combines text-based creation, localized editing and native transparency, with a visual component of 7 billion parameters. This profile identifies the Pro API variant, not the base model's downloadable weights.

Image · Paid proprietary COHERE

Cohere Embed 5 Pro

Embedding model for high-quality indexing of text, images and mixed documents. Supports over one hundred languages and shares a vector space with Embed 5 Fast, allowing documents to be indexed with Pro and queried with Fast without rebuilding the index.

Embeddings · Paid proprietary COHERE

Cohere Embed 5 Fast

Embedding variant for interactive search, agent queries and high traffic. Accepts text, images and mixed documents in over one hundred languages. Shares Pro's vector space to query existing indexes at a lower text cost.

Embeddings · Paid proprietary OPENAI

GPT‑6.1 Sol

OpenAI model for complex coding, computer use and professional work with text and image inputs. The provider describes performance close to Astra at a lower cost; assess that trade-off on your own tasks.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Opus 5.5

Next-generation Opus for agentic coding, computer use, and extended professional work, with a one-million-token context.

Text and reasoning · Paid proprietary ANTHROPIC

Claude Sonnet 5.5

Sonnet update for everyday tasks, code, and documents, with improved speed and a million tokens of context.

Text and reasoning · Paid proprietary XAI

Grok 4.7

xAI model for coding, agents, and knowledge work with text and image input.

Text and reasoning · Paid proprietary XAI

Grok Voice Transcribe 2.0

Second version of xAI’s transcription model for files and live audio.

Voice and audio · Paid proprietary GOOGLE

Gemini 3.8 Live

Live conversation model integrating voice, video, and tools with low latency.

Voice and audio · Paid proprietary GOOGLE

Gemini 3.8 Live Extended Thinking

Live voice variant with background reasoning for complex multi-step tasks.

Voice and audio · Paid proprietary ALIBABA QWEN

Qwen3.8-Max

Qwen3.8 flagship family for software engineering, agents, and visual understanding with a one-million-token context window.

Text and reasoning · Paid proprietary ALIBABA QWEN

Qwen3.8-Flash

Efficient Qwen3.8 variant for text and vision with long context and extended output.

Text and reasoning · Paid proprietary ALIBABA QWEN

Qwen3.8-Omni-Flash

Text, image, audio, and video understanding model with text responses and tools.

Multimodal · Paid proprietary ALIBABA QWEN

Qwen3.8-Omni-Flash-Realtime

Audio-visual model for real-time interaction with voice and text output, tools, and remote MCP.

Voice and audio · Paid proprietary ELEVENLABS

Eleven v4

Expressive speech synthesis model for narration, characters, and dialogue in more than 90 languages.

Voice and audio · Paid proprietary ELEVENLABS

Eleven v4 Turbo

Eleven v4 variant optimized for voice agents and expressive real-time responses.

Voice and audio · Paid proprietary Z.AI

GLM‑5.3

Z.AI open model focused on software engineering, long-running agents, and technical reasoning.

Code · API and open weights Z.AI

GLM‑5.3‑Flash

Multimodal variant with 320.000 billion total parameters and 18.000 billion active parameters, focused on efficiency.

Multimodal · API and open weights ALIBABA QWEN

Qwen Image 3.0 Pro

Vision model for complex compositions, small text, and editing based on up to three reference images.

Image · Paid proprietary ALIBABA QWEN

Wan 3.0 Video Prime

Accelerated version of Wan 3.0 for creating videos using text, image, video, and audio references.

Video · Paid proprietary

Tools per problem

The utility before the infinite directory.

DEVELOPMENT

Code Wizards

Understand repositories, review changes and automate tests.

Code INVESTIGATION

Search with sources

Synthesize evidence while preserving provenance and boundaries.

Evidence CREATION

Image, video, voice and audio

Multimodal generation and editing with control and cost criteria.

Multimedia

Agents and automation

Systems that combine models, tools, memory and control rules.

PATTERN

Agent with tools

Select actions, consult services and verify the result.

Tools PATTERN

Monitored flow

Reserve sensitive decisions for explicit human review.

Control PATTERN

Recovery with sources

Connect answers with documents and traceable references.

RAG

Browse by capacity

Each capability requires different metrics and risks.

TEXT

Text and reasoning

Context, planning, synthesis and problem resolution.

Compare SOFTWARE

Programming

Code generation, understanding and maintenance.

Assess MULTIMODAL

Image, video, voice and audio

Inputs and outputs that need to evaluate fidelity, cost and rights.

Modalities
ANALYSIS

Related analysis

From Data to Verifiable Decision: When to Use AI for Data Analysis and When Not To
GUIA

From Data to Verifiable Decision: When to Use AI for Data Analysis and When Not To

A decision guide for choosing between AI assistants, BI, and code based on the specific analytical task, with verification checklists for data integrity, privacy, and total cost.

17 Oct 2026
Automating operational workflows with AI: when the model becomes the bottleneck
ANALISIS

Automating operational workflows with AI: when the model becomes the bottleneck

A practical framework for deciding when to use rules, predictive models, or generative AI in operational tasks—and what to measure before automating.

17 Oct 2026
Choosing a RAG Tool: Matching Document Type to Error Risk
GUIA

Choosing a RAG Tool: Matching Document Type to Error Risk

A guide to deciding between keyword search, semantic search, and RAG, and how to evaluate document structure, traceability, updates, latency, and cost without confusing relevance with truth.

15 Oct 2026
GLM-5.1: What to Verify Before Deploying It as a Coding Agent
ANALISIS

GLM-5.1: What to Verify Before Deploying It as a Coding Agent

Z.AI presents GLM-5.1 as a model for agentic engineering, and a model card for it is listed on Hugging Face. But the available information is not enough to confirm its license, execution requirements, official pricing, or benchmark results. This guide separates what is documented from what remains unverified and offers a checklist for evaluating the model.

03 Oct 2026
Cohere Embed 5 Pro: what its retrieval advantage demonstrates—and what remains to be known before choosing it
ANALISIS

Cohere Embed 5 Pro: what its retrieval advantage demonstrates—and what remains to be known before choosing it

Cohere says Embed 5 Pro achieved the highest average score among the models it tested and highlights its performance on financial document retrieval. The available information does not make it possible to verify the score, methodology, exact price, or all conditions of use.

03 Oct 2026
Cohere Embed 5 Fast: What We Know About Its Capabilities, Access, and Pricing
ANALISIS

Cohere Embed 5 Fast: What We Know About Its Capabilities, Access, and Pricing

The available documentation identifies Embed 5 Fast as part of the Embed 5 family, but it is not enough to confirm its input modalities, availability by channel, applicable price, or retrieval results. Here is what can be stated—and what still needs to be verified.

03 Oct 2026