Account and models
BYOK model catalog (September 2026)
This page mirrors the **Compare** view at `/dashboard/ai/compare`. Catalog refreshed **2026-09-01**.
This page mirrors the Compare view at /dashboard/ai/compare. Catalog refreshed 2026-09-01.
Best performance
For peak DeplAI performance—security analysis, deploy pipelines, and multi-step agents—use:
| Model | Provider id | When to use |
|---|---|---|
| MiniMax M3 | MiniMax-M3 | Agentic reasoning, coding, and 1M multimodal context. Set high or extrahigh thinking effort. |
| Grok 4.6 | grok-4.6 | Flagship xAI model for coding and long-running agents. Set high or extrahigh (xhigh) reasoning effort. |
Both models expose a shared effort ladder from low through extrahigh. Reasoning cannot be disabled on Grok 4.6; MiniMax M3 defaults to adaptive thinking.
Platform baseline
DeplAI uses GPT-5.6 Sol (gpt-5.6-sol) as the internal platform baseline for routing and cost accounting. Compare rankings are relative to DeplAI jobs—not a generic chat leaderboard.
Active flagship models
| Display name | Provider | API id | Context | Input / output (USD per 1M tokens) |
|---|---|---|---|---|
| MiniMax M3 | minimax | MiniMax-M3 | 1M | $0.60 / $2.40 |
| Grok 4.6 | xai | grok-4.6 | 500k | $2.00 / $6.00 |
| Grok 4.5 | xai | grok-4.5 | 500k | $2.00 / $6.00 |
| Gemini 3.1 Pro | gemini | gemini-3.1-pro | 1M | $2.00 / $12.00 |
| GPT-5.6 Sol | openai | gpt-5.6-sol | 1.05M | $4.00 / $20.00 |
| GPT-5.6 Terra | openai | gpt-5.6-terra | 1.05M | $2.00 / $12.00 |
| Claude Fable 5 | anthropic | claude-fable-5 | 1M | $10.00 / $50.00 |
| Claude Opus 5 | anthropic | claude-opus-5 | 1M | $5.00 / $25.00 |
| Claude Sonnet 5 | anthropic | claude-sonnet-5 | 1M | $3.00 / $15.00 |
Other active models
| Display name | Provider | API id | Notes |
|---|---|---|---|
| Gemini 3.5 Flash-Lite | gemini | gemini-3.5-flash-lite | Fast, low-cost Gemini tier |
| MiniMax M2.7 | minimax | MiniMax-M2.7 | Prior M-series; always-on reasoning |
| Kimi K2.5 | kimi | kimi-k2.5 | 2M context |
| GLM-5 | glm | glm-5 | Strong coding and agents |
| GPT-5.6 Luna | openai | gpt-5.6-luna | Fast OpenAI tier |
| GPT-5.4 Mini | openai | gpt-5.4-mini | Cost-optimized |
| GPT-5.3 Codex | openai | gpt-5.3-codex | Coding specialist |
| Claude Haiku 4.5 | anthropic | claude-haiku-4-5 | Fast Anthropic tier |
| Llama 3.3 70B (Groq) | groq | llama-3.3-70b-versatile | Hosted inference |
| GPT OSS 120B (Groq) | groq | openai/gpt-oss-120b | Open-weight on Groq |
Deprecated (still visible, not selectable)
| Model | Replacement |
|---|---|
| Gemini 2.5 Pro | Gemini 3.1 Pro |
| Gemini 2.5 Flash | Gemini 3.5 Flash-Lite |
| Claude Opus 4.6 | Claude Opus 5 |
| Claude Sonnet 4.6 | Claude Sonnet 5 |
| Grok 4 | Grok 4.6 |
Thinking effort by provider
| Provider | Parameter | Shared ladder | Default |
|---|---|---|---|
| OpenAI | reasoning.effort | low · medium · high · extrahigh | medium |
| Anthropic | output_config.effort | low · medium · high · extrahigh | high |
| xAI | reasoning_effort | low · medium · high · extrahigh | high |
| Gemini | thinking_level | low · medium · high | medium |
| MiniMax | thinking.type (M3) | low · medium · high · extrahigh | medium |
OpenAI and Anthropic also expose extended API values (max, none, product-only ultra). xAI maps extrahigh to API xhigh.
Logical aliases
Agents can request logical names instead of a provider id:
| Alias | Resolves to capability |
|---|---|
best / best_reasoning | Highest reasoning score among your keyed providers |
best_coding | Strongest coding model |
best_agent | Best agentic / tool-use fit |
best_fast | Fast latency profile |
best_cost | Lowest output price |
best_long_context | Largest context window |
best_multimodal | Vision + audio where supported |
With keys for MiniMax and xAI, best_reasoning typically ranks MiniMax M3 or Grok 4.6 first—still use high or extrahigh effort for peak results.
Related: Core concepts · Security and data · Billing