GPT-3.5 Turbo Instruct pricing
Source: litellm · updated 2026-11-17
Short answer: GPT-3.5 Turbo Instruct costs $1.50 per million input tokens and $2.00 per million output tokens. That is a 3:1 blended rate of $1.62 per million tokens, ranking it #636 of 823 tracked models by input price.
Input / 1M
$1.50
Output / 1M
$2.00
Output ÷ input
1.33×
Context window
8,192
Price breakdown
| Metric | USD |
|---|---|
| Input, per 1M tokens | $1.50 |
| Output, per 1M tokens | $2.00 |
| Output ÷ input multiple | 1.33× |
| 3:1 blended, per 1M tokens | $1.62 |
| Context window | 8,192 tokens |
| Vendor | vercel_ai_gateway |
What a typical workload costs
| Monthly volume | Input cost | Output cost | Total |
|---|---|---|---|
| 1M in / 250K out | $1.50 | $0.500 | $2.00 |
| 10M in / 2.5M out | $15.00 | $5.00 | $20.00 |
| 100M in / 25M out | $150 | $50.00 | $200 |
Other vercel_ai_gateway models
| Model | Input | Output |
|---|---|---|
| Ministral 3 B | $0.040 | $0.040 |
| Llama 3.8 B | $0.050 | $0.080 |
| Devstral Small | $0.070 | $0.280 |
| Qwen 3 14 B | $0.080 | $0.240 |
| Qwen 3 30 B | $0.100 | $0.300 |
Cheaper alternatives
Next cheaper option: Anthropic Claude 3.5 Haiku at $0.800 per 1M input tokens. Next step up: Gemini 3.5 Flash at $0.750 per 1M input tokens.
Spotted a price that is out of date? Submit a correction.