GLM-5.3

GLM-5.3 is a deliberate controlled experiment. Z.ai kept the base model, the architecture, and both parameter counts identical to GLM-5.2, then spent a month scaling post-training alone.

The stack underneath is the one GLM-5.2 introduced. The config declares GlmMoeDsaForCausalLM: 78 layers, 256 routed experts plus 1 shared expert, 8 experts active per token, and a hidden size of 6,144. IndexShare handles long-context efficiency by reusing one sparse-attention indexer across each group of four layers, selecting the top 2,048 tokens per query, which keeps the KV cache affordable at 1M tokens. SAO carries the reinforcement learning on long-horizon tasks. Training runs on slime, the open source RL framework from Z.ai, with Megatron on the training side and SGLang on rollout.

What changed is scale and plumbing:

• More environments, not more parameters. Z.ai built pipelines that synthesize long-horizon task environments end to end, along with the reward signal for a subset. A judge agent attempts each task to confirm it is solvable, and verifiers are written without access to the reference solution.
• A faster training loop. Workload-aware scheduling and a local-storage caching layer lifted end-to-end RL throughput by more than 2.3x on long-horizon coding tasks. Training and rollout log probabilities were aligned to the 1e-7 level.
• Fewer tokens per task. The efficiency gain is the underrated part. At max effort GLM-5.3 reaches 34.5% on the in-house Z.ai Code Bench using about 75K output tokens per task, where GLM-5.2 needed 96K to reach 23.4%.

The released checkpoint is FP8 throughout, E4M3 with 128x128 block scaling, spread across 141 safetensors shards for roughly 756GB. Z.ai shipped no BF16 version, so quantization is part of the release rather than something the community adds later. At about one byte per parameter the weights land on a single 8-GPU node, which is what makes a 753B model practical to self-host at all.

For more models like this one, browse the full open source LLM directory, compare the field in the best open source LLMs in 2026, or read the open source LLM ecosystem statistics.

GLM-5.3 ships at the same API rates as GLM-5.2: $1.40 per million input tokens, $0.26 per million cached input tokens, and $4.40 per million output tokens. A capability jump at a flat price is the headline here. Z.ai lists cached input storage as free for a limited time. Since August 28, 2026 the API is no longer the only route, because the open weights make self-hosting a real alternative for anyone with an 8-GPU node.