--- title: "Together AI API changes, deprecations and deadlines | Harbyn" description: "Together AI: 0 upcoming deadlines and 25 recorded changes as of 2026-10-25, each with its source." canonical: "https://harbyn.com/vendors/together" last-updated: 2026-10-25 --- # Together AI: API changes and deadlines Markdown version of . AI models and APIs. 0 upcoming deadlines, 25 recorded changes. Every entry cites the Together AI page we read it from. ## What we read The Together AI pages and feeds our ingestion visits, politely and on a schedule. - **Changelog**: - **Deprecations**: - **Changelog**: ## What the scanner looks for How `harbyn scan` recognises Together AI in a repository, on your machine. Nothing leaves it. - **Packages**: `npm together-ai`, `pypi together` - **API hosts**: `api.together.xyz`, `api.together.ai` ## Upcoming deadlines No upcoming Together AI deadline in the feed right now. ## Change history Newest first, by the date Together AI announced it. This is the archive: past changes are not warnings, and changes we recorded after they took effect are marked as such. ### October 2026 - Oct 18, 2026, Feature: [DeepSeek-V4.1-Flash-1017 available on serverless](https://docs.together.ai/docs/changelog.md) - Oct 15, 2026, Feature: [Muse Spark Open 816B and Nemotron 3.5 Super VL available on serverless](https://docs.together.ai/docs/changelog.md) - Oct 13, 2026, Feature: [Post-training for custom training loops on Together GPUs](https://docs.together.ai/docs/changelog.md) - Oct 2, 2026, Rate limit: [serverless no longer applies dynamic rate limits](https://docs.together.ai/docs/changelog.md) - Oct 2, 2026, Feature: [New provisioned throughput models](https://docs.together.ai/docs/changelog.md) - Oct 1, 2026, Feature: [Code sandbox SDK and CLI](https://docs.together.ai/docs/changelog/rss.xml) - Oct 1, 2026, Feature: [Together Link beta](https://docs.together.ai/docs/changelog.md) ### September 2026 - Sep 29, 2026, Behavior change: [Batch API files retained for 7 days](https://harbyn.com/changes/together/2026-09-29-batch-api-files-retained-for-7-days.md) - Sep 29, 2026, Feature: [Longer fine-tuning context for Qwen 27B models](https://docs.together.ai/docs/changelog/rss.xml) - Sep 29, 2026, Behavior change: [Qwen 27B fine-tuning batch size limits changed](https://harbyn.com/changes/together/2026-09-29-qwen-27b-fine-tuning-batch-size-limits-changed.md) - Sep 29, 2026, Behavior change: [LoRA rank limit raised to 128, older SDKs now default to rank 128](https://harbyn.com/changes/together/2026-09-29-september-29-2026.md) - Sep 24, 2026, Feature: [deploy a fine-tuned model by its registry name](https://docs.together.ai/docs/changelog/rss.xml) - Sep 24, 2026, Notice: [TogetherLink beta](https://docs.together.ai/docs/changelog) (retracted) - Sep 23, 2026, Notice: [Tev1-4B-experimental available on serverless](https://docs.together.ai/docs/changelog) - Sep 22, 2026, Pricing: [lower serverless pricing for Qwen3.7-Max and Qwen3.8-Flash](https://harbyn.com/changes/together/2026-09-22-qwen-serverless-price-cut.md) (effective Sep 22, 2026; recorded after the fact) - Sep 16, 2026, Feature: [Automatic idle shutdown for dedicated deployments](https://docs.together.ai/docs/changelog) - Sep 15, 2026, Feature: [Rollouts for dedicated endpoints and CLI v2.34.0](https://docs.together.ai/docs/changelog/rss.xml) - Sep 14, 2026, Feature: [MiniMaxAI/MiniMax-H3 available on serverless](https://docs.together.ai/docs/changelog) - Sep 11, 2026, Notice: [DeepSeek-V4.1-Flash available on serverless](https://docs.together.ai/docs/changelog) - Sep 10, 2026, Behavior change: [Together CLI v2.33.2 improvements](https://harbyn.com/changes/together/2026-09-10-preemptible-compute-for-gpu-clusters-preemptible-compute-is-now-in-public-pr-2.md) - Sep 10, 2026, Breaking change: [Preemptible GPU compute preview; CLI v2.33.2 removes --scale-to-zero-window](https://harbyn.com/changes/together/2026-09-10-preemptible-gpu-compute-and-cli-v2-33-2.md) - Sep 8, 2026, Notice: [zai-org/GLM-5.3 available for fine-tuning](https://docs.together.ai/docs/changelog) - Sep 1, 2026, Retirement: [four serverless models removed on September 14, 2026](https://harbyn.com/changes/together/2026-09-01-four-serverless-models-removed-september-14-2026.md) (effective Sep 14, 2026; recorded after the fact) - Sep 1, 2026, Pricing: [H100 80GB dedicated endpoint price cut to $3.99 per hour](https://harbyn.com/changes/together/2026-09-01-h100-dedicated-endpoint-price-cut.md) ### August 2026 - Aug 29, 2026, Feature: [active\_sessions autoscaling metric for TTS dedicated endpoints](https://docs.together.ai/docs/changelog/rss.xml)