Topic
The stage after pretraining where a model is adapted via fine-tuning, reinforcement learning, or reward shaping to produce specific behaviors or skills.
CoreWeave has announced Forge, a platform linking model training, inference and evaluation so production results improve models and agents. Buyers comparing AI clouds now have a tooling claim to test next to chip capacity.
Reality
NVIDIA says fine-tuned Nemotron 3 models reached gold-medal level at IMO 2026 and IOI 2026, scoring 30 of 42 at the officially graded IMO. The company credits standard fine-tuning paired with a loop that generates and checks answers.
Reality
Tenet is post-trained on Moonshot's Kimi K3, converting per-call payments to OpenAI, Anthropic and Google into a fixed training bill. The counterparty risk moved rather than disappeared.
Reality
Google DeepMind's Koray Kavukcuoglu says Gemini 4 is in post-training and under test in Antigravity, with release hoped for well before the end of 2026. With no benchmarks published, the first evidence teams get will come from their own agent tasks.
Reality
Two agentic post-training jobs, a 1.02-trillion-parameter model and a 309-billion cousin, ran with step time, reward and infrastructure error rates on the open web. Both stopped near step 30, and both models are still unreleased.
Reality
Koa is a post-trained Nemotron 3 Super running inside Salesforce's trust boundary, taught by reinforcement learning on simulated enterprise workflows. The case the company makes for it rests on consistency and inspectable provenance.
Reality
Turing Post talked through agent feedback with Salesforce's chief AI scientist at Dreamforce. A correction lands in one of four places: the context, persistent memory, the software around the model, or the weights.
Reality
The $15.5 billion price works out near 39 times annual recurring revenue above $400 million, and it rests on a margin claim Harvey has not quantified: that tuning Kimi K3 beats paying OpenAI per call.
Reality
The $15.5bn valuation is 38.75 times the $400m of annual recurring revenue Harvey disclosed, and the in-house model meant to defend that price is post-trained on a Beijing lab's open weights rather than on backer OpenAI's.
Reality
A Tsinghua-led study handed model training to coding agents and got real gains out of them, then found the agents almost never revised the method they chose in the first few minutes, which is the exact part that successor-building roadmaps assume.
Reality