|
| 1 | +--- |
| 2 | +title: How Copilot serves AI models |
| 3 | +intro: 'Learn how {% data variables.copilot.copilot_chat_short %} serves different AI models.' |
| 4 | +versions: |
| 5 | + feature: copilot |
| 6 | +topics: |
| 7 | + - Copilot |
| 8 | +--- |
| 9 | + |
| 10 | +{% data variables.product.prodname_copilot %} can use a variety of AI models. This article explains how {% data variables.product.prodname_copilot_short %} serves these models. |
| 11 | + |
| 12 | +## OpenAI models |
| 13 | + |
| 14 | +Used for: |
| 15 | + |
| 16 | +* {% data variables.copilot.copilot_o3 %} |
| 17 | +* {% data variables.copilot.copilot_o4_mini %} |
| 18 | + |
| 19 | +These models are hosted by OpenAI and {% data variables.product.github %}'s Azure tenant when used in {% data variables.product.prodname_copilot %}. OpenAI makes the [following data commitment](https://openai.com/enterprise-privacy/): _We [OpenAI] do not train our models on your business data by default_. {% data variables.product.github %} maintains a [zero data retention agreement](https://platform.openai.com/docs/guides/your-data) with OpenAI. |
| 20 | + |
| 21 | +When using OpenAI's models, input requests and output responses continue to run through {% data variables.product.prodname_copilot %}'s content filters for public code matching, when applied, along with those for harmful or offensive content. |
| 22 | + |
| 23 | +## Anthropic models |
| 24 | + |
| 25 | +Used for: |
| 26 | + |
| 27 | +* {% data variables.copilot.copilot_claude_opus %} |
| 28 | +* {% data variables.copilot.copilot_claude_sonnet_35 %} |
| 29 | +* {% data variables.copilot.copilot_claude_sonnet_37 %} |
| 30 | +* {% data variables.copilot.copilot_claude_sonnet_37 %} Thinking |
| 31 | +* {% data variables.copilot.copilot_claude_sonnet_40 %} |
| 32 | + |
| 33 | +{% data variables.copilot.copilot_claude_opus %} and {% data variables.copilot.copilot_claude_sonnet_40 %} are hosted by Anthropic PBC and Google Cloud Platform. {% data variables.copilot.copilot_claude_sonnet_37 %} is hosted by Amazon Web Services, Anthropic PBC, and Google Cloud Platform. {% data variables.copilot.copilot_claude_sonnet_35 %} is hosted exclusively by Amazon Web Services. {% data variables.product.github %} has provider agreements in place to ensure data is not used for training. Additional details for each provider are included below: |
| 34 | + |
| 35 | +* Amazon Bedrock: Amazon makes the [following data commitments](https://docs.aws.amazon.com/bedrock/latest/userguide/data-protection.html): _Amazon Bedrock doesn't store or log your prompts and completions. Amazon Bedrock doesn't use your prompts and completions to train any AWS models and doesn't distribute them to third parties_. |
| 36 | +* Anthropic PBC: {% data variables.product.github %} maintains a [zero data retention agreement](https://privacy.anthropic.com/en/articles/8956058-i-have-a-zero-retention-agreement-with-anthropic-what-products-does-it-apply-to) with Anthropic. |
| 37 | +* Google Cloud: [Google commits to not training on {% data variables.product.github %} data as part of their service terms](https://cloud.google.com/vertex-ai/generative-ai/docs/data-governance). {% data variables.product.github %} is additionally not subject to prompt logging for abuse monitoring. |
| 38 | + |
| 39 | +To provide better service quality and reduce latency, {% data variables.product.github %} uses [prompt caching](https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching). You can read more about prompt caching on [Anthropic PBC](https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching), [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/prompt-caching.html), and [Google Cloud](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude-prompt-caching). |
| 40 | + |
| 41 | +When using {% data variables.copilot.copilot_claude %}, input prompts and output completions continue to run through {% data variables.product.prodname_copilot %}'s content filters for public code matching, when applied, along with those for harmful or offensive content. |
| 42 | + |
| 43 | +## Google models |
| 44 | + |
| 45 | +Used for: |
| 46 | + |
| 47 | +* {% data variables.copilot.copilot_gemini_25_pro %} |
| 48 | +* {% data variables.copilot.copilot_gemini_flash %} |
| 49 | + |
| 50 | +{% data variables.product.prodname_copilot %} uses {% data variables.copilot.copilot_gemini_flash %} and {% data variables.copilot.copilot_gemini_25_pro %} hosted on Google Cloud Platform (GCP). When using {% data variables.copilot.copilot_gemini %} models, prompts and metadata are sent to GCP, which makes the [following data commitment](https://cloud.google.com/vertex-ai/generative-ai/docs/data-governance): _{% data variables.copilot.copilot_gemini %} doesn't use your prompts, or its responses, as data to train its models._ |
| 51 | + |
| 52 | +When using {% data variables.copilot.copilot_gemini %} models, input prompts and output completions continue to run through {% data variables.product.prodname_copilot %}'s content filters for public code matching, when applied, along with those for harmful or offensive content. |
0 commit comments