Sep 29 and Oct 9, 2026 Model Deprecation Notice - Documentation

The following models will be retired from the Novita platform in two batches, on Sep 29 and Oct 9, 2026. If your application uses an affected model, migrate to the recommended replacement model before its retirement date. A — in the replacement column means the model has no direct replacement.

Affected models

Model ID	Retirement date	Replacement model
`deepseek/deepseek-v3.2`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`deepseek/deepseek-r1`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`deepseek/deepseek-v3.1`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`deepseek/deepseek-v3.1-terminus`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`deepseek/deepseek-v3.2-exp`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`deepseek/deepseek-r1-distill-llama-70b`	Sep 29, 2026	—
`deepseek/deepseek-v3/community`	Sep 29, 2026	`deepseek/deepseek-v4.1-flash`
`qwen/qwen3.6-plus`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-vl-235b-a22b-instruct`	Oct 9, 2026	`qwen/qwen3.8-max`
`openai/gpt-oss-20b`	Oct 9, 2026	—
`deepseek/deepseek-v4-flash-vision-exp`	Oct 9, 2026	`zai-org/glm-5.3-flash`
`qwen/qwen3-coder-30b-a3b-instruct`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-coder-480b-a35b-instruct`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-coder-next`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen-2.5-72b-instruct`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-next-80b-a3b-instruct`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3.5-plus`	Oct 9, 2026	`qwen/qwen3.8-max`
`sao10k/l31-70b-euryale-v2.2`	Oct 9, 2026	—
`qwen/qwen3-vl-235b-a22b-thinking`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-max`	Oct 9, 2026	`qwen/qwen3.8-max`
`mistralai/mistral-nemo`	Oct 9, 2026	—
`sao10k/l3-8b-lunaris`	Oct 9, 2026	—
`zai-org/glm-5v-turbo`	Oct 9, 2026	`zai-org/glm-5.3-flash`
`qwen/qwen3-235b-a22b-thinking-2507`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-235b-a22b-instruct-2507`	Oct 9, 2026	`qwen/qwen3.8-max`
`baidu/ernie-4.5-vl-424b-a47b`	Oct 9, 2026	—
`minimax/minimax-m2`	Oct 9, 2026	`minimax/minimax-m3`
`qwen/qwen3-omni-30b-a3b-instruct`	Oct 9, 2026	`qwen/qwen3.8-flash`
`minimax/minimax-m2.1`	Oct 9, 2026	`minimax/minimax-m3`
`zai-org/glm-5-turbo`	Oct 9, 2026	`zai-org/glm-5.3-flash`
`minimax/minimax-m2.5-highspeed`	Oct 9, 2026	`minimax/minimax-m3`
`qwen/qwen3-235b-a22b-fp8`	Oct 9, 2026	`qwen/qwen3.8-max`
`qwen/qwen3-embedding-0.6b`	Oct 9, 2026	—
`thudm/glm-4-32b-0414`	Oct 9, 2026	`zai-org/glm-5.3-flash`

Action required

If your application uses an affected model, migrate to the recommended replacement model before its retirement date. For models with no direct replacement, remove the affected calls from your application before that date.

Last modified on October 10, 2026

⌘I