Today, we're introducing AC2, the Applied Compute Agent Cloud, to enable every team to train, serve, and improve their own frontier models.

The adoption of open weights is accelerating, but how do we securely deploy them? We've assembled a framework for assessing and mitigating risks in open weight model deployments based on our experiences post-training with frontier enterprises.

Gap analysis is how our Applied Researchers, like @_brylee10, surface agent failures across billions of tokens of RL traces for our customers. Using our platform, AC2, and low-cost classifiers like Jev, we can catch 85% of failure modes at a fraction of the cost of LLM judges and turn them into training data.

we partnered with @appliedcompute to post-train a small model for large-scale code search over precomputed indexes at 300 repos, this is ~3x faster than using filesystem + grep, and reduces the marginal cost of a search by up to 100x vs frontier models turbopuffer.com/blog/large-s…

Out of the box the base model is a weak searcher. Over training, correctness x citation support improves 57% on the narrow task and precision improves 211% on the open-ended one. The agent also learns where to look. ripgrep falls from 15.3 calls per rollout to 0.3 as turbopuffer search takes over.

Agents are the new discovery layer. We've seen that across 2M+ conversations with Ask DoorDash, where people want agents to step in and carry their intent through to a fulfillable order.

“50% of DoorDash’s agentic restaurant orders are going to places users have never ordered from before.” @andyfang tells our CEO @ypatil125 what happens when agents become the discovery layer. If models increasingly decide what gets surfaced and bought, companies have a strong reason to train and own that intelligence.

“Until you actually see things operationally, it’s going to be hard to build DoorDash from scratch.” Our CEO @ypatil125 sat down with @andyfang on why cheaper software doesn’t erase years of operating advantage. @DoorDash’s moat is its proprietary data, edge cases, and hard-won knowledge, and increasingly, the models trained on top of it.

AFAIK AC2 Is the only platform with support for full weight fine tuning of Kimi K3. For short, simple tasks LoRA fine tuning may be similar. But for multi turn, agentic workloads, full weight training is strictly superior. dm if you’re interested in trying AC2

We are excited to support GLM 5.3 for training and inference on the Applied Compute Platform. Post-train it on your data, run inference at scale, and keep the weights.

It's interesting to see the recent capabilities jump in smaller models like DeepSeek-V4-Flash-0731, Qwen3.8-27B, and now GLM-5.3-Flash. As the baseline capabilities of small models continue to improve, we get a much more compelling starting point for model specialization. The post-trained performance will push the global frontier, and the cost and efficiency wins become more clear. I'm also hopeful that in the long run, capable small models will unlock on-device possibilities and custom post-training at a personal level.

We expect GLM-5.3-Flash to be one of the most popular models to post-train with AC2. DM for access.