6 releases so far from to
Everything Together AI shipped in November 2026, collected from official Together AI sources. Each entry is dated and links to the page it came from.
Post-training: GLM-5.3 Flash, Messages API and Responses API Support, New serverless models, and Post-training: Kimi K3.1 Pro were released. GLM-5.3 Flash added LoRA training with extended context. API support expanded compatibility. New serverless models included moonshotai/Kimi-K3.1-Pro with 1M context length. Kimi K3.1 Pro offered post-training options.
Summary generated from the releases listed below.
November 2026 is Together AI's 4th busiest month of the last 12: 6 releases so far, close to its monthly average of 7. That is 20 releases fewer than October 2026, with the month still in progress. Most were Model (3), followed by Developer Tools (2) and API (1).
Releases so far
6
-20 vs October 2026
Rank of 12 months
4th
Average 7.1 a month
Most shipped
Model
3 releases
The following models are now available on serverless: moonshotai/Kimi-K3.1-Pro: 1,048,576 context length, NVFP4 quantization, text and image input, function calling, and structured outputs. Pricing: $2.00 input / $10.00 output / $0.20 cached input (per 1M tokens).
Source: Together AI Changelogmoonshotai/Kimi-K3.1-Pro is available for post-training: LoRA training on B200-SXM, a 163,840-token training context, a 167,936-token generator context, rank up to 32, and up to three adapter sessions per resource. You can’t deploy its adapters on dedicated model inference yet.…
Source: Together AI ChangelogLoRA training on B200-SXM with 159,744-token training context and 163,840-token generator context
Source: Together AI ChangelogNewest first. Each entry links to the official page it came from.
The following models are now available on serverless: moonshotai/Kimi-K3.1-Pro: 1,048,576 context length, NVFP4 quantization, text and image input, function calling, and structured outputs. Pricing: $2.00 input / $10.00 output / $0.20 cached input (per 1M tokens).
moonshotai/Kimi-K3.1-Pro is available for post-training: LoRA training on B200-SXM, a 163,840-token training context, a 167,936-token generator context, rank up to 32, and up to three adapter sessions per resource. You can’t deploy its adapters on dedicated model inference yet.…
Together Python SDK 3.1.0 and Together CLI 3.1.0 are now available. tg beta jig deploy no longer stores your API key as a TOGETHER_API_KEY secret on new deployments. Existing deployments keep the secret they already have. Sprocket 0.1.8 and later authenticate with the deployment…
Together AI now supports the Anthropic Messages API at POST /v1/messages and the OpenAI Responses API at POST /v1/responses. Both APIs support zai-org/GLM-5.3 and zai-org/GLM-5.2 on serverless inference, and zai-org/GLM-5.3-Flash and zai-org/GLM-5.2 on dedicated model inference.
Post-training: Prepare MiniMax M3.1 Flash adapter for FP4 inference on dedicated model inference. Base model ships in serving format; job merges adapter (1 hour).
LoRA training on B200-SXM with 159,744-token training context and 163,840-token generator context
Together AI releases per month, for November 2026 and the 11 months before it. A taller bar means more releases that month; the dated entries for each month are on its own page.
10 of the companies we track shipped in November 2026; Together AI ranks 1st by number of releases.
Newsletter
One monthly digest of the sharpest launches, model drops, and product moves from every company we track, delivered in one clear read.
The standout launches, model drops, and platform moves worth tracking.
One tight read instead of bouncing between company blogs and timelines.
Clear summaries you can scan fast and forward to the team when needed.
Monthly dispatch
Get a clean monthly dispatch covering the launches that mattered, not every minor product breadcrumb.
One email each month, tuned for signal over noise.
