AI models for text and analysis. Overview of 16 providers and systems (June 2026).

AI Models for Text, Data Analysis, and Content

A practical overview of 16 AI providers and systems (as of June 2026). Filter by group, search by name, or sort alphabetically. Click a card to see details, strengths/weaknesses, and pricing.

← All AI model catalogs Image-generation models →

AI Tools

4 tools to choose a model, compare providers, craft a prompt, and calculate ROI.

Wizard

Model Selection Wizard

Not sure which model to choose? Answer 9 questions and get a personalized recommendation.

  • Budget, compliance, integrations
  • Recommendation with % match score
  • Cost estimate and rationale
Get started Generator

AI Prompt Creator

Generate precise prompts for ChatGPT, Claude, Gemini, and more. 8 steps to the perfect instruction.

  • Step-by-step analysis, Few-shot, Tree of Thought
  • 13 ready-made templates
  • "Improve my prompt" mode + quality score
Create prompt Calculator

AI Implementation ROI Calculator

Calculate real return on investment in AI. Concrete numbers: savings, costs, and break-even.

  • 10 task types × 11 industries
  • ROI%, net profit, break-even
  • Downloadable and copyable report
Calculate ROI Compare

AI Model Comparison

Pick 2-4 providers and see a side-by-side view: who it's for, when to use, costs, strengths, and weaknesses.

  • 16 providers in one view
  • Quick presets (frontier, open-weight)
  • Data from the AI model catalog
Compare

AI Models Catalog

Filter and compare models by group, name, and estimated costs.

Frontier models (API)
Claude (Anthropic)
Sonnet 5 - agentic Sonnet near Opus 4.8, 1M context
Claude Sonnet 5 (released 06/30/2026) is the most agentic Sonnet model yet - planning, tools (browser, terminal) and autonomous work at a level that recently required larger, more expensive models.
Sonnet 5: promo $2 input / $10 output per 1M tok. through 08/31/2026, then $3/$15; Opus 4.8: $5/$25; Fable 5: $10/$50; long-context pricing above 200k input. Details →
Frontier models (API)
OpenAI (GPT)
GPT-5.5 in ChatGPT; GPT-Live voice; GPT-5.6 API preview
GPT-5.5 (Pro/Instant, Apr-May 2026) remains the default text model in ChatGPT: agentic capabilities, 1M+ context, native multimodality and NVIDIA co-design.
Paid API per 1M tokens; GPT-5.6 preview: Sol $5/$30, Terra $2.50/$15, Luna $1/$6; details in OpenAI documentation. Details →
Frontier models (API)
Google Gemini
3.5 Flash, 3.1 Pro, Vertex AI, Zero Data Retention
Gemini 3.1 Pro (02/19/2026) and Gemini 3.5 Flash (05/19/2026, Google I/O) are natively multimodal models (text, audio, image, video); Gemini 3.5 Pro announced.
Paid API (Gemini API / Vertex); competitive with existing GCP infrastructure. Details →
Frontier models (API)
Grok (xAI)
Grok 4.3, 1M context, real-time X data
xAI develops the Grok family with up to 1M tokens in the frontier Grok-4.3 variant.
Pay-as-you-go: Grok-4.3 $1.25/$2.50; Grok-4.1 Fast $0.20/$0.50; cache read significantly below input rate. Details →
Office ecosystems
Microsoft Copilot
GPT-5.5 in M365, multi-model routing, enterprise
Copilot is a family of AI assistants in Microsoft products: Copilot for 365 (~$30/user/mo) with GPT-5.5 and multi-model routing (including Claude Opus 4.8) built into Outlook, Excel, Teams and PowerPoint; GitHub Copilot (~$10-19/user/mo) for code in VS Code and JetBrains; Azure OpenAI Service for custom deployments.
Copilot for 365: ~$30/user/mo; GitHub Copilot: ~$10-19/user/mo; Azure OpenAI: pay-as-you-go. Details →
Research & search
Perplexity
search with citations, frontier models in Pro
Perplexity combines a search engine with an LLM - answers with a list of sources and citations (grounded generation).
Free / Pro: $20/mo / Max: $200/mo; Enterprise Pro: $40/user/mo; Enterprise Max: $325/user/mo. Details →
Open-weight & self-host
Mistral AI
Medium 3.5 and Small 4 open-weight; European provider; GPAI Code
Mistral develops its open-weight line: Medium 3.5 and Small 4 (Mar-Apr 2026) alongside the Mistral 3 family and Pixtral 12B.
Medium 3.5 API: $1.50 input / $7.50 output per 1M tokens; self-hosting open weights costs GPU infrastructure. Details →
Open-weight & self-host
DeepSeek
V4-Pro/Flash, DSA, MIT, legacy sunset
DeepSeek V4-Pro and V4-Flash (04/24/2026) replace earlier V3.x variants: Dynamic Sparse Attention (DSA) architecture, MIT license, high reasoning and code performance.
V4 Flash/Pro: $0.14/$0.435 input and $0.28/$0.87 output per 1M tokens; MIT open-weight = GPU cost. Details →
Open-weight & self-host
Qwen (Alibaba Cloud)
Open 3.5/3.6 + Qwen3.7 Max/Plus (API, 1M); Qwen3.5-Plus as cheaper API tier
Qwen3.5 and smaller Qwen3.6 variants (open-weight, Apache 2.0) are the current 'workhorse' series in the open ecosystem: MoE at various scales, VLM (e.g.
Low-medium: open-weight = infrastructure; Qwen3.7-Max API: $2.50/$7.50 per 1M tokens; Qwen3.7-Plus from $0.40/$1.60; Qwen3.5-Plus from $0.40/$2.40. Details →
Open-weight & self-host
GLM (Z.ai)
GLM-5.2: 1M context and agents; GLM-4.7-Flash for lighter self-hosting
GLM is a model family from Z.ai, formerly known as Zhipu AI.
Z.ai pricing on June 23, 2026: GLM-5.2 costs $1.40 input, $0.26 cached input and $4.40 output per 1M tokens; GLM-4.7-Flash is listed as free in the API. Self-hosting adds GPU, power and operating costs. Details →
Open-weight & self-host
Llama (Meta)
Scout 10M ctx, Maverick 1M, MoE 17B active
Llama 4 Scout/Maverick (04/05/2025): open-weight MoE family (17B active) on 'Community License' (not OSI): >700M MAU clause, AUP, restrictions on using outputs to improve other LLMs.
Very flexible: weights 'free', costs on the deployment side; hosted (Groq, Together etc.) - low to medium. Details →
Specialist & agentic
Kimi (Moonshot AI)
K2.7 Code + K2.6, 256K ctx, multimodal
Kimi K2.7 Code is Moonshot AI's strongest coding variant; Kimi K2.6 remains the general model with a 256K context window and multimodality (text+image+video).
API: K2.7 Code and K2.6 - $0.95 input / $0.19-0.16 cached input / $4.00 output per 1M tokens; K2.5 legacy - $0.60 / $3.00. Details →
Specialist & agentic
Manus
autonomous AI agent; Meta acquisition blocked (NDRC)
Manus (public since 03.2025) is an autonomous agentic platform - plans, executes and delivers results of multi-step tasks without supervision.
Standard: $20/mo (4000 cr.); Custom: $40/mo (8000 cr.); Extended: $200/mo (40000 cr.); free tier available. Details →
Specialist & agentic
Sakana Fugu
multi-model orchestration through one API; Fugu and Fugu Ultra
Sakana Fugu, commercially released on June 22, 2026, is not a conventional standalone LLM.
Subscriptions: $20 / $100 / $200 monthly; Fugu Ultra PAYG: $5 input, $30 output and $0.50 cached input per 1M tokens, with higher rates above 272K context. Details →
Polish language models
Bielik (SpeakLeash)
best Polish open-weight quality, Apache 2.0
Bielik is a family of Polish language models developed by SpeakLeash and ACK Cyfronet AGH on Polish HPC infrastructure (Helios, Athena).
Open-weight (Apache 2.0): cost = GPU infrastructure; available on Hugging Face. Details →
Polish language models
PLLuM
Polish 4B-70B family; base, instruct and chat models
PLLuM (Polish Large Language Model) is a model family developed by a consortium of Polish research and public institutions.
Public model weights; cost depends on model size, GPU hosting and operations. Check the license of the exact repository before deployment. Details →

This overview is approximate (June 2026) and does not replace official documentation or provider pricing. Before making a purchase decision, check the current terms on provider websites.