How AI Model Pricing Works

Understanding token costs and why different models charge different amounts

💡 Key Concept

1 Internal Token = €0.01

We convert OpenAI's actual costs into "tokens" for transparent billing. Different AI models have different prices per OpenAI token, so the same conversation might use different amounts of OpenAI tokens but cost the same in our internal tokens.

Real Example: Same Question, Different Models

gpt-5-nano (Cheapest)

OpenAI Tokens Used:793 tokens

Input Cost:$0.05 per 1M tokens

Output Cost:$0.40 per 1M tokens

Actual Cost:$0.000142

You Pay:1 token (€0.01)

gpt-4.1-mini (Balanced)

OpenAI Tokens Used:259 tokens

Input Cost:$0.40 per 1M tokens

Output Cost:$1.60 per 1M tokens

Actual Cost:$0.000234

You Pay:1 token (€0.01)

📊 Notice: gpt-5-nano used MORE OpenAI tokens (793 vs 259) but cost LESS in dollars ($0.000142 vs $0.000234) because it's cheaper per token. Both round up to 1 internal token (€0.01) for billing.

Why Do Models Use Different Token Counts?

Cheaper Models (Nano, Mini)

✓Generate longer, more verbose answers (use more tokens)
✓May repeat concepts or use simpler language
✓Each token costs very little ($0.05-$0.40 per million)

Premium Models (GPT-5, GPT-4o)

✓Generate concise, efficient answers (use fewer tokens)
✓Better compression and understanding
✓Each token costs more ($1.25-$10.00 per million)

Complete Model Pricing

Model	Input (per 1M)	Output (per 1M)	Best For
gpt-5.1	$1.25	$10.00	Latest GPT-5 flagship model with best reasoning and quality.
gpt-5	$1.25	$10.00	GPT-5 base model with excellent performance.
gpt-5-mini	$0.25	$2.00	GPT-5 Mini - excellent balance of cost and quality.
gpt-5-nano	$0.05	$0.40	Smallest GPT-5 model - ultra cost-effective for simple tasks.
gpt-4.1	$2.00	$8.00	GPT-4.1 with improved reasoning over GPT-4.
gpt-4.1-mini	$0.40	$1.60	GPT-4.1 Mini - balanced performance and cost.
gpt-4.1-nano	$0.10	$0.40	GPT-4.1 Nano - cheapest option for basic tasks.
gpt-4o-mini	$0.15	$0.60	Default for text generation. Fast and cost-effective.
gpt-4o	$2.50	$10.00	Premium quality with better reasoning and creativity.

* Pricing is from OpenAI. We charge based on actual usage converted to internal tokens (1 token = €0.01).

Audio Models Pricing

Diligentify supports multiple voice interaction modes. Each uses different audio models with different pricing.

⚠️ Important: Voice Costs More

Voice modes use significantly more tokens than text-based practice questions.

Text Practice Questions

Cost per Question:1-2 tokens

2,000 tokens gets you:~1,000 questions

✓ Best value for practice

Realtime Voice Sessions

Cost per 15 min:~400 tokens

2,000 tokens gets you:~5 sessions

✓ Immersive live practice

💡 Pro Tip: Use text-based practice for high-volume training, and save Realtime voice sessions for final mock interviews or when you need live conversation practice.

Realtime Models (Live Voice)

Most expensive - charges for audio input, audio output, and text processing separately.

Model	Audio Input	Audio Output	Text
gpt-4o-realtime-preview-2024-12-17	$1000/1M tokens	$4000/1M tokens	Current stable version. Most tested and reliable.
gpt-realtime(Latest)	$1000/1M tokens	$4000/1M tokens	2025 version with improved natural expression and responsiveness.
gpt-4o-mini-realtime-preview-2024-12-17	$500/1M tokens	$2000/1M tokens	Cost-effective option. Exact pricing TBD by OpenAI.

* Realtime models charge separately for audio input (your voice), audio output (AI voice), and text processing.

Speech-to-Text (Transcription)

Model	Cost
gpt-4o-transcribe	$0.006 per minute
gpt-4o-transcribe-diarize	$0.006 per minute
gpt-4o-mini-transcribe	$0.003 per minute(Cheapest!)
whisper-1	$0.006 per minute

Text-to-Speech (TTS)

Model	Cost
gpt-4o-mini-tts	$0.015 per minute
tts-1	$15 per 1M characters
tts-1-hd	$30 per 1M characters