Model Tracker
GPT-3.5 Turbo
GPT-3.5 Turbo is an AI model made by OpenAI. It is deprecated. It takes text and returns text. Its context window is 16,385 tokens.
GPT-3.5 Turbo is an AI model made by OpenAI. It is deprecated. It takes text and returns text. Its context window is 16,385 tokens. Its longest answer is 4,096 tokens. Through the API it costs $0.50 per 1M tokens for input and $1.50 per 1M tokens for output. Its knowledge cutoff is 1 September 2021.
- Released
- Not confirmed: The maker's announcement says the model is being released but does not print its original release date on the page; the page displays only an update date.
- Status
- Deprecated — developers.openai.com, read 16 October 2026
- Takes
- Text — developers.openai.com, read 16 October 2026
- Returns
- Text — developers.openai.com, read 16 October 2026
- Open weights
- Not confirmed: The maker's model page does not state whether GPT-3.5 Turbo's weights are open or downloadable.
- Licence
- Not confirmed: The maker's model page does not name a licence for GPT-3.5 Turbo's weights.
- Context window
- 16,385 tokens — developers.openai.com, read 16 October 2026
- Longest answer
- 4,096 tokens — developers.openai.com, read 16 October 2026
- Knowledge cutoff
- 1 September 2021 — developers.openai.com, read 16 October 2026
- API price, input
- $0.50 per 1M tokens — developers.openai.com, read 16 October 2026
- API price, output
- $1.50 per 1M tokens — developers.openai.com, read 16 October 2026
- Where
- API — developers.openai.com, read 16 October 2026
- Parameters
- Not confirmed: The maker's model page does not publish a parameter count for GPT-3.5 Turbo.
- Model card
- Not confirmed: two readings disagree: https://developers.openai.com/api/docs/models/gpt-3.5-turbo and https://platform.openai.com/docs/models/gpt-3.5-turbo
Benchmarks
Measured by others, not by the maker: each score with the name the measurer gives the model and the setting it ran, its day and its licence.
Epoch AI
- Epoch Capabilities Index, GPT-3.5 Turbo (Nov 2023): 118.6 (109.7–121.8), as of 19 October 2026
- Epoch Capabilities Index, GPT-3.5 Turbo (Jan 2024): 115.7 (108.0–119.4), as of 19 October 2026
- Epoch Capabilities Index, GPT-3.5 Turbo (Jun 2023): 113.4 (104.8–118.0), as of 19 October 2026
- GPQA Diamond, gpt-3.5-turbo-1106: 28.0% (± 1.9), run 27 January 2025
- GPQA Diamond, gpt-3.5-turbo-0125: 27.2% (± 2.0), run 27 January 2025
- FrontierMath, tiers 1–3, gpt-3.5-turbo-0125: 0.0% (± 0.0), run 27 August 2026
- OTIS Mock AIME 2024–2025, gpt-3.5-turbo-0125: 2.2% (± 2.2), run 7 August 2026
Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks, CC BY 4.0: Epoch AI’s own runs and its Capabilities Index. Data · CC BY 4.0
Questions
Is GPT-3.5 Turbo still available?
It is deprecated: OpenAI has announced the date it will retire it.
What is the context window of GPT-3.5 Turbo?
The context window of GPT-3.5 Turbo is 16,385 tokens, as OpenAI prints it.
How much does the GPT-3.5 Turbo API cost?
On the standard tier, GPT-3.5 Turbo costs $0.50 per 1M tokens for input and $1.50 per 1M tokens for output, as OpenAI lists it.
What does GPT-3.5 Turbo take and return?
GPT-3.5 Turbo takes text and returns text.
What is the knowledge cutoff of GPT-3.5 Turbo?
The knowledge cutoff of GPT-3.5 Turbo is 1 September 2021.
More from OpenAI
- ada
- babbage
- babbage-002
- code-davinci-002
- computer-use-preview
- curie
- dall-e-3
- DALL·E
- DALL·E 2
- davinci
- davinci-002
- GPT-3.5 Turbo Instruct
- GPT-4
- GPT-4 Turbo
- GPT-4.1
- GPT-4.1 mini
- GPT-4.1 nano
- GPT-4.5
- GPT-4.5 Preview
- GPT-4o
- GPT-4o mini
- GPT-4o Mini Search Preview
- GPT-4o Mini Transcribe
- GPT-4o Mini TTS
Sources
- GPT-3.5 Turbo Model | OpenAI API (read 16 October 2026)
- Introducing APIs for GPT-3.5 Turbo and Whisper (read 16 October 2026)