> ## Documentation Index
> Fetch the complete documentation index at: https://help.wonka.chat/llms.txt
> Use this file to discover all available pages before exploring further.
# Model comparison
> Every model in WonkaChat side by side: what it is best at, speed, cost, region and public benchmarks.
WonkaChat gives you 14 AI models from 6 providers. This page helps you pick one: start with the task you have in mind, then compare the models in detail.
Not sure? Pick the **Router** in the model picker: it chooses a model for each message. See [Choosing a model](/en/improved-usage/choosing-a-model).
## Which model for which task
| Task | Recommended models |
| - | - |
| Everyday writing, emails and summaries | [GPT-5.6 Luna](#model-gpt-5-6-luna), [Claude Sonnet 5.5](#model-claude-sonnet-5-5), [Mistral Large 4](#model-mistral-large-4) |
| Connected business tools (ERP, CRM, email) | [GPT-6.1 Sol](#model-gpt-6-1-sol), [Claude Sonnet 5.5](#model-claude-sonnet-5-5), [GPT-5.6 Luna](#model-gpt-5-6-luna) |
| Complex analysis, strategy and hard problems | [Claude Opus 5.5](#model-claude-opus-5-5), [GPT-6.1 Sol](#model-gpt-6-1-sol), [Claude Sonnet 5.5](#model-claude-sonnet-5-5) |
| Quick answers and high volumes | [Claude Haiku 5.5](#model-claude-haiku-5-5), [Gemini 3.8 Flash](#model-gemini-3-8-flash), [Gemini 3.1 Flash Lite](#model-gemini-3-1-flash-lite) |
| Images, scans and visual documents | [Gemini 3.8 Flash](#model-gemini-3-8-flash), [Gemini 3.1 Pro (Vertex AI)](#model-gemini-3-1-pro--vertex-ai), [GPT-6.1 Sol](#model-gpt-6-1-sol) |
| Very long documents | [Gemini 3.1 Pro (Vertex AI)](#model-gemini-3-1-pro--vertex-ai), [Kimi K3](#model-kimi-k3), [GLM 5.2](#model-glm-5-2) |
| Data that must stay in Europe | [Mistral Large 4](#model-mistral-large-4), [GPT-5.6 Luna](#model-gpt-5-6-luna), [Claude Sonnet 5.5](#model-claude-sonnet-5-5) |
## Recommended models
These models are highlighted in the model picker. They cover most needs.
OpenAI Β· β¬ Β· πͺπΊ EU
Balanced everyday work at a low cost
OpenAI Β· β¬β¬ Β· πͺπΊ EU
Strong reasoning for demanding professional work and connected tools
Mistral AI Β· β¬ Β· πͺπΊ EU
European model for capable, cost-efficient work
Anthropic Β· β¬ Β· πͺπΊ EU
Fast answers for high volumes and cost-sensitive work
Anthropic Β· β¬β¬β¬ Β· πͺπΊ EU
The most capable Claude model, for the hardest tasks
## All models
| Model | Provider | Best for | Speed | Cost | Region |
| - | - | - | - | - | - |
| **GPT-5.6 Luna** | OpenAI | Balanced everyday work at a low cost | βββββ | β¬ | πͺπΊ EU |
| **GPT-6.1 Sol** | OpenAI | Strong reasoning for demanding professional work and connected tools | βββββ | β¬β¬ | πͺπΊ EU |
| **Mistral Large 4** | Mistral AI | European model for capable, cost-efficient work | βββββ | β¬ | πͺπΊ EU |
| **Claude Haiku 5.5** | Anthropic | Fast answers for high volumes and cost-sensitive work | βββββ | β¬ | πͺπΊ EU |
| **Claude Opus 5.5** | Anthropic | The most capable Claude model, for the hardest tasks | βββββ | β¬β¬β¬ | πͺπΊ EU |
| **Claude Sonnet 5.5** | Anthropic | Strong reasoning with balanced speed and natural writing | βββββ | β¬β¬ | πͺπΊ EU |
| **Gemini 3.1 Flash Lite** | Google | The lowest-cost option for light tasks | βββββ | β¬ | πͺπΊ EU |
| **Gemini 3.1 Pro (Vertex AI)** | Google | Advanced reasoning on long and visual documents (preview) | βββββ | β¬β¬β¬ | π Global |
| **Gemini 3.7 Flash** | Google | Previous fast Gemini model for text, images and documents | βββββ | β¬β¬β¬ | πͺπΊ EU |
| **Gemini 3.8 Flash** | Google | Fast work on text, images and documents | βββββ | β¬β¬β¬ | πͺπΊ EU |
| **Mistral Medium** | Mistral AI | Efficient European model for everyday work | βββββ | β¬β¬ | πͺπΊ EU |
| **Kimi K2.5** | Moonshot AI | Efficient work on long documents | βββββ | β¬β¬ | π Global |
| **Kimi K3** | Moonshot AI | Reasoning on very long documents | βββββ | β¬β¬β¬β¬ | π Global |
| **GLM 5.2** | Z.ai | Reasoning on long documents, hosted in Europe | βββββ | β¬ | πͺπΊ EU |
## Benchmarks and technical details
**Intelligence Index** comes from [Artificial Analysis Intelligence Index v4.3.2](https://artificialanalysis.ai/leaderboards/models), retrieved on October 15, 2026. A higher score is better. It measures general capability on public tests, not how well a model works with your own tools and data.
| Model | Intelligence Index | Reasoning | Context window |
| - | - | - | - |
| **Claude Opus 5.5** | 54 | Yes | 1M tokens |
| **GPT-6.1 Sol** | 50 | Yes | 1M tokens |
| **Claude Sonnet 5.5** | 47 | Yes | 1M tokens |
| **Kimi K3** | 44 | Yes | 1M tokens |
| **Gemini 3.8 Flash** | 41 | Yes | 1M tokens |
| **Gemini 3.7 Flash** | 39 | No | 1M tokens |
| **Mistral Large 4** | 38 | Yes | 1M tokens |
| **Claude Haiku 5.5** | 38 | Yes | 1M tokens |
| **GLM 5.2** | 34 | Yes | 1M tokens |
| **GPT-5.6 Luna** | 32 | Yes | 1M tokens |
| **Gemini 3.1 Pro (Vertex AI)** | 30 | Yes | 1M tokens |
| **Kimi K2.5** | 23 | No | 262K tokens |
| **Gemini 3.1 Flash Lite** | 16 | No | 1M tokens |
| **Mistral Medium** | 14 | No | 131K tokens |
## How to read this page
* **Speed**: how fast the model answers, from β (slow) to βββββ (very fast).
* **Cost**: how many credits the model uses compared with the others, from β¬ (lowest) to β¬β¬β¬β¬ (highest). See [AI usage and credits](/en/pricing/tokens-guide) for how credits work.
* **Region**: where the model runs. πͺπΊ EU stays in the European Union, πΊπΈ US runs in the United States, π Global can run in any region of the provider.
* **Reasoning**: the model thinks step by step before answering. Needed for connected business tools (ERP, CRM, email).
* **Context window**: how much text the model can read at once (1M tokens is roughly 750,000 words).
Your administrator decides which models your organization can use. If a model listed here is missing from your model picker, ask them.
*Model list updated on October 15, 2026.*
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.