Northwind Aerial Services · signed in as Avery Morgan (owner)
Each AI step needs a model of its kind (text, image, video, voice, transcription). Usage is charged in credits, so a dearer model uses more of the plan’s credits.
| Module | Step | Kind | Model | Credits per run |
|---|---|---|---|---|
| Assistant | Assistant answer | text | 47 | |
| Assistant | Assistant reads an image | image | Claude Sonnet 5.5 (vision) | 5 |
| Fleet | Fleet insight digest | text | 12 | |
| Fleet | Footage review (2 min clip) | video | Gemini (video understanding) | 240 |
| Missions | Go / no-go advice | text | 24 | |
| Maintenance | Maintenance forecast | text | 30 | |
| Maintenance | Inspection photo check | image | Claude Sonnet 5.5 (vision) | 5 |
| Safety | Shift briefing | text | 21 | |
| Safety | Spoken shift briefing | voice | ElevenLabs Multilingual (speech) | 216 |
| Compliance | Monthly report draft | text | 120 | |
| Compliance | Voice note transcription (5 min) | transcription | Whisper (transcription) | 30 |
1 credit = $0.001 of model cost at catalogue prices. The prices behind the credits are estimates that have not been checked against the providers yet.
| Model | Provider | Kind | Used for | Price status |
|---|---|---|---|---|
| Claude Sonnet 5.5default | Anthropic | text | Insights, advisor, forecasts, briefings and reports | Estimate, not verified · source |
| Claude Haiku 5.5 | Anthropic | text | Cheaper, faster text for frequent digests | Estimate, not verified · source |
| Claude Opus 5.5 | Anthropic | text | Hardest reasoning: incident analysis and monthly reports | Estimate, not verified · source |
| GPT-5.1 | OpenAI | text | OpenAI flagship for harder insights and monthly reports | Estimate, not verified · source |
| GPT-5 mini | OpenAI | text | Faster OpenAI digests and routine answers | Estimate, not verified · source |
| GPT-5 nano | OpenAI | text | Short OpenAI classification and frequent checks | Estimate, not verified · source |
| Grok 4.7 | xAI | text | xAI flagship. List rate is for prompts under 200K tokens | Estimate, not verified · source |
| Grok 4.3 | xAI | text | Fast xAI answers. List rate is for prompts under 200K tokens | Estimate, not verified · source |
| GLM 5.2 (open weights) | Z.ai | text | Open weights. Frontier open model for reports and advice | Estimate, not verified · source |
| Nemotron 3 Ultra (open weights) | NVIDIA | text | Open weights. Large NVIDIA model for digests and advice | Estimate, not verified · source |
| Nemotron 3 Nano (open weights) | NVIDIA | text | Open weights. Cheap NVIDIA model for frequent digests | Estimate, not verified · source |
| Nemotron 3.5 Super VL (open weights) | NVIDIA | text | Open weights. NVIDIA vision-language model, priced per token | Estimate, not verified · source |
| Claude Sonnet 5.5 (vision)default | Anthropic | image | Reading inspection photos: cracks, hot spots, damage | Estimate, not verified · source |
| Gemini (video understanding)default | video | Reviewing drone footage for events | Estimate, not verified · source | |
| ElevenLabs Multilingual (speech)default | ElevenLabs | voice | Spoken shift briefings for pilots in the field | Estimate, not verified · source |
| Whisper (transcription)default | OpenAI | transcription | Transcribing pilot voice notes and radio calls | Estimate, not verified · source |