Supported AI Models
OpenAI GPT-5, Anthropic Claude 4, Google Gemini, xAI Grok — which plans include which models
Supported AI Models
Deslicer supports multiple AI providers. You select the model per agent. Free plans have limited model access; Standard and Enterprise include all models.
Navigation
- Parent: Reference
- Related: Credits and Usage | Creating Agents
Available Models
OpenAI
| Model | Category | Description |
|---|---|---|
| GPT-5.5 | Flagship | Frontier reasoning and agentic performance (1M+ context) |
| GPT-5.4 | Flagship | General-purpose flagship with web search and audio |
| GPT-5.4 Nano | Efficient | Lightweight variant for fast, low-cost tasks |
| GPT-5.3 Chat | Flagship | Chat-optimized snapshot |
| GPT-5.3 Codex | Flagship | Specialized for coding workflows |
| GPT-5.2 | Flagship | Strong coding and agentic task model |
| GPT-5 | Flagship | Built-in thinking with web search and audio |
| GPT-4o Mini | Efficient | Fast and affordable multimodal model (Free tier) |
Anthropic
| Model | Category | Description |
|---|---|---|
| Claude 4.7 Opus | Flagship | Frontier reasoning and agentic flagship (default model, 1M context) |
| Claude 4.6 Opus | Flagship | Strong reasoning with 1M context (beta) |
| Claude 4.6 Sonnet | Flagship | Frontier Sonnet for agents, coding, and long context |
| Claude 4.5 Opus | Flagship | Capable model with reduced pricing |
| Claude 4 Opus | Flagship | Sustained coding performance |
| Claude 4 Sonnet | Flagship | Precise instruction following |
| Claude 3.7 Sonnet | Flagship | High capability Sonnet |
| Claude 3.5 Sonnet | Flagship | Balanced speed and intelligence |
| Claude 3.5 Haiku | Efficient | Fast and cost-effective |
| Model | Category | Description |
|---|---|---|
| Gemini 3 Pro (Preview) | Flagship | Latest Gemini Pro preview (1M context) |
| Gemini 3 Flash (Preview) | Efficient | Fast Gemini 3 preview |
| Gemini 2.5 Pro | Flagship | Enhanced reasoning with web search (1M context) |
| Gemini 2.5 Flash | Efficient | Fast model with web search |
| Gemini 2.0 Flash | Efficient | Capable model with web search (1M context) |
| Gemini 2.0 Flash Lite | Efficient | Lightweight Gemini for fast responses |
Each model has different strengths. Claude 4.7 Opus is the default model and is well-suited for complex Splunk tasks. Try different models per agent to find the best fit.
Web Search Support
When you enable web search on an agent, Deslicer uses provider-native web search tools. The following providers offer built-in web search capabilities:
| Provider | Web Search | Notes |
|---|---|---|
| OpenAI | Supported | All GPT-5 and GPT-4o models |
| Anthropic | Supported | Claude 4+ Opus and Sonnet models |
| Supported | Gemini 2.0+ models | |
| xAI | Supported | Grok models (when enabled) |
If you enable web search on an agent, the model automatically gains the ability to look up current information from the internet during conversations. Web search works reliably across all supported providers, with automatic fallback handling for quota limits and transient errors. See Chatting with Agents for details.
Plan Access
| Plan | Model Access |
|---|---|
| Free | GPT-4o Mini (additional models may be limited) |
| Standard | All models |
| Enterprise | All models |
Upgrade to Standard or Enterprise for full model access. See Plans and Pricing for details.
Selecting Models per Agent
When you create or edit an agent, you choose the model. The same agent always uses that model unless you change it. Different agents can use different models — e.g., Search Ninja on Claude 4.7 Opus and BOTS Hunter on GPT-5.
Credits are consumed based on model and token usage. Higher-capability models typically cost more per message. Track usage in Settings → Billing → Usage.
