Snowflake Cortex Fine-Tuning
SnowflakeSnowflake Cortex Fine-Tuning is a fully managed function inside Snowflake Cortex AI that applies parameter-efficient fine-tuning (PEFT) to a short, fixed list of supported base models, such as Llama 3.1 8B. Training data comes from tables in the customer's account, and both training and inference run on Snowflake compute rather than a third-party platform, billed per token in credits.
Openness
1 high confidence- license
- Proprietary
- service
- fully-managed-proprietary(in-Snowflake)
- source
- closed
- method
- PEFT(LoRA-class, not specified)
- base-models
- fixed allowlist(narrowed to a single model, llama3.1-8b, per the current page)
- access
- Snowflake-account+credits
- data
- user-data stays in account but engine closed
A fully managed, proprietary fine-tuning feature inside Snowflake, with no source published and no way to self-host. The base-model allowlist is a single model, llama3.1-8b.
- https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-finetuning recorded 2026-09-23
Cortex Fine-tuning is a fully managed service that runs entirely inside Snowflake via the FINETUNE SQL function or the Snowsight AI Studio wizard; the page names no published repository, no self-host option, and no license terms for the underlying engine. The base-models table now lists a single row, "llama3.1-8b", with the expand/collapse controls disabled, confirming the table is complete; the page still notes models available for fine-tuning may be added or removed in the future.
Adoption
not assessedSnowflake does not publish how many customers use this fine-tuning feature. Its documentation only describes a view for tracking an account's own usage, so there is no honest way to judge how widely it is used.
- https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-finetuning recorded 2026-09-28
Documents the "CORTEX_FINE_TUNING_USAGE_HISTORY view" for tracking an account's own consumption; searched the body for customers, customer count, deployments and adoption figures: none found.
Capability
2 medium confidenceSnowflake Cortex Fine-Tuning offers a single parameter-efficient training method on a single documented base model, Llama 3.1 8B, with no full fine-tuning and no reinforcement or preference-based training. That makes it a convenience feature for Snowflake users, well short of managed services such as OpenAI's and Together's that add preference and reinforcement tuning on many more models.
- https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-finetuning recorded 2026-09-28
Describes "parameter-efficient fine-tuning (PEFT) to create customized adaptors" and a base-models table listing only "llama3.1-8b"; searched the body for LoRA, RLHF, DPO, full fine-tuning and mixtral: none found.
Verified 2026-09-23