GLM-5.3
- In Input modalities: Text Out Output modalities: Text
- Reasoning· low, high, max
- Tool use
- Structured output
- Open weights
Prices
4 offerings · prices as listedglm-5.3| Conditions/ 1M tokens | Input | Cache read | Output |
|---|---|---|---|
| Standard | $1.40 | $0.26 | $4.40 |
Prices from Pricing (opens in a new tab) (checked Nov 8, 2026)
glm-5.3| Conditions/ 1M tokens | Input | Cache read | Output |
|---|---|---|---|
| Standard | ¥8.00 | ¥2.00 | ¥28.00 |
Prices from 智谱 BigModel：API 定价 (opens in a new tab) (checked Nov 9, 2026)
Standard: the price a request with no conditions pays (the headline price elsewhere on the site).
Details
- Family
- glm
- Model ID
Not stated on Zhipu AI’s pages: max input, max thinking budget, built-in tools, knowledge cutoff, licence, aliases.
Run locally
Weights: zai-org/GLM-5.3- Parameters 753.3B
- Active 51.2B (estimated) · 256 experts, 8 per token
- Native precision FP8
- Layers 78
- KV cache per token 88 KB
- Config positions 1024K (config.json max_position_embeddings)
- API context 977K (maker’s API)
- Needs
- ≈ 392.0 GB (weights 372.7 + KV cache 0.7 + overhead 18.6)
- Usable
- 21.6 GB (90% of its memory)
- Try
- it’s too large for this hardware; see the smallest that runs it below
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| FP8 (native)maker’s file | 703.7 GB | ≈ 739.6 | ✕Doesn’t fit | — |
| Q6_K | 576.2 GB | ≈ 605.7 | ✕Doesn’t fit | — |
| Q5_K_M | 499.0 GB | ≈ 524.6 | ✕Doesn’t fit | — |
| Q5_0 | 482.3 GB | ≈ 507.1 | ✕Doesn’t fit | — |
| Q4_K_M | 426.2 GB | ≈ 448.2 | ✕Doesn’t fit | — |
| Q4_0 | 394.6 GB | ≈ 415.1 | ✕Doesn’t fit | — |
| IQ4_XS | 372.7 GB | ≈ 392.0 | ✕Doesn’t fit | — |
| Q3_K_Mquality drops | 342.9 GB | ≈ 360.7 | ✕Doesn’t fit | — |
| IQ3_XXSquality drops | 268.4 GB | ≈ 282.5 | ✕Doesn’t fit | — |
| Q2_Kquality drops | 263.1 GB | ≈ 276.9 | ✕Doesn’t fit | — |
| IQ2_XXSquality drops | 180.7 GB | ≈ 190.4 | ✕Doesn’t fit | — |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): no single listed machine; about 7 × 80 GB GPUs.
Run commands
The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-11. How it’s estimated
Cited scores
From third-party leaderboards, not measured by us- Epoch Capabilities Index 156.9
#29 of 118 listed models · CI 154.8–159.2
- LMArena Text 1,483
#20 of 163 listed models · max effort · CI 1,478–1,487 · 30,524 votes
- Epoch AI · Chess Puzzles 37%
#37 of 85 listed models · max effort · ±4.85
- Epoch AI · FrontierMath Tier 4 v2 29.3%
#33 of 65 listed models · max effort · ±7.19
- Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Oct 28 – Nov 5, 2026
- LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Nov 8, 2026
- EQ-Bench by Samuel J. Paech (eqbench.com; EQ-bench/EQ-bench-site on GitHub), MIT License. · License · as of Oct 28, 2026
- LiveBench (White et al.), release 2026-06-25, livebench.ai, licensed CC BY-SA 4.0. This board, adapted from it, is shared under CC BY-SA 4.0. · License · as of Jun 25, 2026
- Terminal-Bench 4.0 leaderboard (Marten, Shaw et al., Terminal-Bench; harbor-framework/terminal-bench on GitHub), licensed Apache-2.0. · License · as of Sep 29, 2026
- changes we made
Provider-reported scores
Published by Zhipu AI, under the conditions it states| Benchmark | Setting | Score |
|---|---|---|
| Terminal-Bench 2.1 | Claude Code 2.1.207 harness, 6h timeout | 88.2% |
| Humanity's Last Exam | with tools, context management, 300K max context; subset not stated | 62.5% |
Included in coding plans
- 5× the usage of the Lite plan
- GLM Coding Plan Lite $18.00 / month2,000 credits every 5 hours10,000 credits per week
- GLM Coding Plan Pro $80.00 / month12,000 credits every 5 hours60,000 credits per week
- GLM Coding Plan Max $168.00 / month28,000 credits every 5 hours140,000 credits per week
- TRAE Free Free500 credits per month
- TRAE Pro ¥99.00 / month4,000 credits per month
News
All news →- Aug 18, 2026Release Zhipu AI releases GLM-5.3 1M-token context window. On Z.ai API: $1.40 input, $4.40 output per 1M tokens.
Changes
All changes →- Now priced on Cloudflare Workers AI Input $1.40 · Cache read $0.26 · Output $4.40
- New language model
Sources
- zai-org/GLM-5.3 model card (opens in a new tab) huggingface.co · checked Nov 8, 2026
- Pricing (opens in a new tab) docs.z.ai · checked Nov 8, 2026
- 硅基流动：模型价格in Chinese (opens in a new tab) siliconflow.cn · checked Nov 8, 2026
- 智谱 BigModel：API 定价in Chinese (opens in a new tab) docs.bigmodel.cn · checked Nov 9, 2026
- GLM-5.3 (opens in a new tab) docs.z.ai · checked Nov 8, 2026
- 智谱 BigModel：模型概览in Chinese (opens in a new tab) docs.bigmodel.cn · checked Nov 4, 2026
- New Released (opens in a new tab) docs.z.ai · checked Nov 8, 2026
- zai-org/GLM-5.3 model card (opens in a new tab) huggingface.co · checked Nov 8, 2026
- Z.ai GLM-5.3 overview (opens in a new tab) docs.z.ai · checked Nov 8, 2026
- GLM-5.3 (opens in a new tab) docs.z.ai · checked Nov 9, 2026
- Cloudflare Workers AI pricing (opens in a new tab) developers.cloudflare.com · checked Nov 7, 2026
- zai-org/GLM-5.3 model card (opens in a new tab) huggingface.co · checked Nov 8, 2026
- zai-org/GLM-5.3 config.json (opens in a new tab) huggingface.co · checked Nov 8, 2026
- zai-org/GLM-5.3 files (opens in a new tab) huggingface.co · checked Nov 7, 2026
Zhipu AI’s official policies: data use, privacy, terms →
Zhipu AI · China · z.ai
Related models
Snapshots, the same family and other models from this provider