1 deadline our listings state. These are quoted dates, not additional change records.
Together also removed these models from serverless inference in 2026, so requests for them fail: openai/gpt-oss-20b on 2026-09-15 (Together recommends Qwen/Qwen3.5-9B); google/gemma-4-31B-it on 2026-09-15 (recommended: zai-org/GLM-5.3-Flash); zai-org/GLM-5.1 on 2026-07-10; Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8 on 2026-06-04; deepseek-ai/DeepSeek-R1 on 2026-05-14; deepseek-ai/DeepSeek-V3.1 on 2026-05-14; meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 on 2026-03-31; meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo on 2026-03-06; meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo on 2026-02-25; meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo on 2026-02-06; meta-llama/Llama-4-Scout-17B-16E-Instruct on 2026-02-06.
Source ↗