Default behavior
If you don’t have a preference, leave the model set to (no preference). This is the default, and it means Abundly will use whatever model we think works best for general agentic behavior. We continuously evaluate models and update this default as better options become available. This is the right choice for most users—you get great performance without having to think about model selection at all.Model aliases vs specific versions
When selecting a model, you can choose between:- Aliases like “Claude Sonnet Latest” — automatically points to the latest version that we have verified works well
- Specific versions like “Claude Sonnet 4.5” — locked to that exact version
Available models
Abundly integrates with the leading AI providers. You can switch models at any time—even mid-conversation.Claude
Anthropic’s models, known for nuanced reasoning and safety
GPT
OpenAI’s models, versatile and widely capable
Gemini
Google’s models, strong at knowledge tasks and multimodal understanding
Comparing models
Not sure which model to choose? We tend to default to the Claude models (Sonnet or Opus). We find they work very well for agentic behaviour. But the other models have improved a lot lately, and some have specific strenghts that you may want to leverage. This is a changing landscape, so you’re probably best off researching online yourself (or asking an Agent to do it for you…). But here’s a high level overview of how the models from each provider compare against each other.Anthropic Claude models
| Model | Best for | Speed | Cost | Notes |
|---|---|---|---|---|
| Claude Opus 4.5 | Complex reasoning, high-stakes tasks | Moderate | $$$ | Most capable Claude model. Best for difficult problems that defeat Sonnet. |
| Claude Sonnet 4.5 | Everyday work, coding, analysis | Fast | $$ | Recommended default. Excellent balance of capability and speed. |
| Claude Haiku 4.5 | High-volume, speed-critical tasks | Fastest | $ | ~3x faster than Sonnet at 1/3 the cost. Near-Sonnet quality for most tasks. |
Google Gemini models
| Model | Best for | Speed | Cost | Notes |
|---|---|---|---|---|
| Gemini Pro 3 | Maximum reasoning depth | Moderate | $$$ | Strongest reasoning. 1M token context for huge documents. |
| Gemini Flash 2.5 | Balanced performance | Fast | $$ | Good multimodal capabilities. Native audio support. |
| Gemini Flash Lite 2.5 | High-volume processing | Fastest | $ | Most cost-effective. Great for bulk summarization and routing. |
OpenAI GPT models
| Model | Best for | Speed | Cost | Notes |
|---|---|---|---|---|
| GPT 5.2 Pro | Mission-critical, complex reasoning | Moderate | $$$$ | Extended reasoning for highest accuracy. Best for high-stakes decisions. |
| GPT 5.2 | Latest improvements, reduced hallucinations | Fast | $$$ | 30% fewer hallucinations. Strong on coding (55.6% SWE-Bench Pro) and reasoning. |
| GPT 5 Mini | Well-defined tasks, interactive apps | Faster | $$ | Great balance of speed and capability. Ideal for chat and coding assistants. |
| GPT 5 Nano | High-volume, simple tasks | Fastest | $ | Ultra-cheap. Best for summarization, classification, and bulk processing. |
Thinking tokens
For Claude models, you can enable thinking—an advanced reasoning mode where the model thinks through problems step by step before responding.
- Thinking for triggers — Enable thinking when the agent responds to emails, schedules, and other triggers
- Default thinking for new chats — Enable thinking by default in chat conversations
- Thinking budget — Control how many tokens the agent can use for thinking (1,024–32,000)
Thinking is currently only available for Claude Sonnet and Opus models.
When to change models
| Scenario | Recommendation |
|---|---|
| Just getting started | Leave it on “(no preference)“ |
| Complex reasoning, planning | Claude Opus, Gemini Pro, or GPT 5.2 Pro |
| Everyday tasks, coding | Claude Sonnet, GPT 5.2, or GPT 5 Mini |
| High-volume, simple tasks | Claude Haiku, Gemini Flash Lite, or GPT 5 Nano |
| Speed is critical | Claude Haiku, Gemini Flash, or GPT 5 Nano |
| Lowest hallucination rates | GPT 5.2 or GPT 5.2 Pro |
| Working with long documents | Gemini Pro (1M token context) |
Unified interface
Regardless of which model you choose, the platform provides a consistent experience:- Switch models even mid-conversation
- Same capabilities work across all models
- Consistent behavior and tool usage
- No need to learn different APIs
What if I pick a model that doesn't work well for my use case?
What if I pick a model that doesn't work well for my use case?
You can switch models at any time. If you’re not happy with the results, try a different model—your agent’s instructions and capabilities stay the same.
How does pricing work for different models?
How does pricing work for different models?
Each model has different credit costs per token. More capable models like Opus and GPT 5.2 cost more per request, while Haiku and Flash Lite are very cost-effective. Check your usage dashboard to monitor credit consumption.
Can I use different models for different agents?
Can I use different models for different agents?
Yes! Each agent can have its own model preference. You might use a fast, cheap model for a high-volume support agent and a powerful reasoning model for a research agent.