# Baseten — The fastest delta weight syncs in the world

- Company: Baseten (baseten.co)
- Announced: 2026-10-30T19:16:58+00:00
- Category: infrastructure-release
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://www.baseten.co/blog/the-fastest-delta-weight-syncs-in-the-world
- Record: https://forck.live/items/29992-the-fastest-delta-weight-syncs-in-the-world
- Subject: Baseten
- Models affected: GLM-5.3, Kimi K3

Baseten has rebuilt its delta weight sync path so that GPU memory holds the only copy of a served policy, cutting a ~770 GB GLM-5.3 sync from 36 seconds to 15 and the request pause from 6 seconds to 0.8. The previous design kept a full parent checkpoint in host RAM, which on an 8×B300 node forced the host KV cache down to 275 GB and added a 45-second CPU promotion per update. The new decoder walks zstd's commands directly, keeps only the nonzero XOR bytes, gathers old values at changed positions, applies the XOR mask and verifies the result against the trainer's checksum before any GPU write. Baseten says the checksum work made the whole update up to 2.3× faster, with Kimi K3 on a 30 GB delta dropping from ~230 to ~100 seconds. Update planning runs in a separate CUDA-free helper process to avoid contending for SGLang's Python event loop, and decoding is pipelined with the download on low-priority native threads sized to leave 16 cores for serving. The change also unblocks Kimi K3 RL rollout on 8:GB300, which Baseten says was previously blocked by CPU memory.

## Evidence

Verbatim from https://www.baseten.co/blog/the-fastest-delta-weight-syncs-in-the-world:

> 770GB of weights previously took ~36 seconds to sync on average; now they take 15, with a sub-second request pause, 7.5x faster than previous SOTA.
> 
> Authors
> 
> Published
> 
> October 30, 2026
> 
> Share
> 
> Published
> 
> October 30, 2026
> 
> In our last technical blog , we showed delta weight syncs for a ~770 GB GLM-5.3 checkpoint in 36 seconds , with 6 seconds of request pause. The same sync now takes 15 seconds, with a 0.8-second pause, enabling throughput improvements by as much as 3.2x.
> 
> We rebuilt the weight sync path around the weights already on the GPU. This removes the need for us to maintain a second full copy in CPU memory. This new design lets us:
> 
> Verify updates by reading only changed GPU positions
> 
> Minimize pause time by preparing GPU writes beforehand
> 
> Use freed host memory for KV cache, improving throughput
> 
> Unlock Kimi K3 RL rollout on 8:GB300, previously blocked by CPU memory
> 
> A quick recap of delta weight syncs
> 
> A trainer in the customer's environment publishes full checkpoints and sparse deltas to object storage. Each delta tensor is zstd(parent_bytes XOR new_bytes) . Once a snapshot is uploaded, the trainer sends a delta weight sync request naming the policy it wants served.

---

Record: https://forck.live/items/29992-the-fastest-delta-weight-syncs-in-the-world
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md