# GLM-5.3-Flash ![Organization](https://img.shields.io/badge/Org-Zhipu%20AI-9333ea) ![Context](https://img.shields.io/badge/Context-1%2C048%2C576%20tokens-1f6feb) ![License](https://img.shields.io/badge/License-MIT-d29922) ![Modality](https://img.shields.io/badge/Modality-Multimodal-8957e5) ![Released](https://img.shields.io/badge/Released-Aug%202026-555) > **Quick answer:** GLM-5.3-Flash is Zhipu AI's (Z.ai) first natively multimodal model in the GLM-5 series — a 320B-total/18B-active-parameter MoE released under the MIT license with a 1,048,576-token (1M) context window. It scores 84.3% on Terminal-Bench 2.1, 63.4% on DeepSWE v1.1, and 1,773 Elo on GDPVal-AA v2, beating its own larger sibling GLM-5.2 across the board at roughly one-tenth the serving cost, while approaching Claude Opus 4.8 on coding and agentic benchmarks. ## At a Glance **Where GLM-5.3-Flash leads** - Open weights under **MIT** — fully self-hostable, unlike the proprietary GLM-5.3 flagship - Outperforms the much larger GLM-5.2 on every benchmark Zhipu reports (Terminal-Bench 2.1 81.0%→84.3%, DeepSWE 46.2%→63.4%, GDPVal-AA v2 1,514→1,773 Elo) at a fraction of the inference cost - First natively multimodal GLM-5 model — hybrid sparse+linear attention architecture keeps long-context serving cheap while adding image/video understanding - Native 1M-token context window (1,048,576 tokens) **Where it lags** - Smaller/cheaper "Flash" tier — the proprietary GLM-5.3 flagship (same generation, API-only) scores higher on every shared benchmark (e.g. Terminal-Bench 2.1 88.2% vs. 84.3%) - Agents' Last Exam (26.3%) trails frontier closed models (Claude Opus 4.8 27.3%, GPT-5.6 Terra 27.0%, Gemini 3.7 Flash 28.0%) by a small but consistent margin - Output is text-only — no image/audio generation **Best for:** Teams that want open, self-hostable multimodal coding/agent capability close to the GLM-5.3/Claude Opus tier without the proprietary-API lock-in or the full flagship's serving cost. ## What GLM-5.3-Flash Is GLM-5.3-Flash is Zhipu AI's (Z.ai) cost-efficient, fully open-weight entry in the GLM-5.3 generation, trained from a newly built base model rather than reusing GLM-5.2's pretrain. It's the first model in the GLM-5 series designed as natively multimodal from the ground up, pairing a 320B-total/18B-active mixture-of-experts language backbone with a dedicated vision tower for image and video understanding. The headline architectural change is a hybrid attention design: most layers use linear attention (Gated DeltaNet-style, "KDA"), with a smaller number of full/sparse ("DeepSeek sparse attention") layers interleaved throughout the 45-layer stack. This sharply reduces the cost of serving very long contexts while preserving retrieval-style long-context accuracy — a departure from the pure full-attention or pure-MoE approach used by earlier GLM releases. The model also adopts Manifold-Constrained Hyper-Connections (mHC), a scaling-efficiency technique layered on top of the standard MoE routing (288 routed experts + 1 shared expert, top-8 routing per token). Zhipu trained it on a fresh 30-trillion-token multimodal pre-training corpus. Before Zhipu's identity was confirmed, this model briefly circulated as an anonymous "stealth" release under third-party branding, generating outsized attention on OpenRouter and community leaderboards purely on capability signal before anyone knew who built it — a reminder that raw benchmark performance travels faster than provenance in this market. Now that it's confirmed as GLM-5.3-Flash with a real org, MIT license, and full HF model card, it stands as one of the strongest fully open-weight multimodal releases of August 2026. ## Specifications | Field | Value | |---|---| | Organization | Zhipu AI (Z.ai) | | Parameters | 320B total / 18B active (MoE) | | Context window | 1,048,576 tokens (1M) | | Architecture | Hybrid linear + sparse attention MoE, 45 layers, 288 routed + 1 shared experts (top-8), Manifold-Constrained Hyper-Connections (mHC), dedicated vision encoder (image + video) | | License | MIT | | Release date | August 2026 | | Modality | Text, image, video input; text output | ## Pricing Open weights (MIT) — free to download and self-host; hosting/inference cost only. Also available hosted via the [Z.ai API](https://docs.z.ai/guides/llm/glm-5.3-flash) — check Z.ai's pricing page for current per-token rates (not independently verified here). ## Context Window GLM-5.3-Flash has a **1,048,576-token (1M) context window** — enough to hold entire large codebases or hundreds of pages of documents in a single request. The hybrid linear/sparse attention design specifically targets keeping this scale of context affordable to serve, rather than treating it as a stretch-goal extension. ## Public Benchmark Scores | Benchmark | Score | Source | Date | |---|---|---|---| | [TerminalBench 2.1](/benchmarks/tbench/terminalbench-21) | 84.3% | Z.ai / HF model card | 2026-08 | | [DeepSWE](/benchmarks/datacurve/deepswe) (v1.1) | 63.4% | Z.ai / HF model card | 2026-08 | | [Agents' Last Exam](/benchmarks/agents-last-exam) | 26.3% | Z.ai / HF model card | 2026-08 | | [Humanity's Last Exam](/benchmarks/cais/humanitys-last-exam) (w/ tools) | 55.3% | Z.ai / HF model card | 2026-08 | | [GDPVal-AA v2](/benchmarks/artificial-analysis/gdpval-aa-v2) | 1,773 Elo | Z.ai / HF model card | 2026-08 | *Scores above are reported by Zhipu AI/third parties and shown for context. They are not Benchgen measurements. Figures depend on harness, tools, and effort settings (default `reasoning_effort=max`) — see the source for methodology.* **Additional benchmark reported by Zhipu, not yet on Benchgen's leaderboard:** - **AutomationBench v1.0.6**: 48.8% — this version includes a bug fix (null-type handling, PR #13) that measurably raises scores versus Benchgen's existing AutomationBench entries for other models (e.g. DeepSeek V4 Pro 31.8%, Kimi K3 30.8%, Qwen3.8 Max 27.3% on the pre-fix version) — the same version-mismatch reason the sibling [GLM-5.3](/models/zhipuai/glm-5-3) page also excludes this score. Not added pending version reconciliation. ## GLM-5.3-Flash vs Alternatives | Model | Terminal-Bench 2.1 | DeepSWE | Agents' Last Exam | GDPVal-AA v2 | |---|---|---|---|---| | **GLM-5.3-Flash** | 84.3% | 63.4% | 26.3% | 1,773 | | [GLM-5.3](/models/zhipuai/glm-5-3) (flagship, proprietary) | 88.2% | 66.9% | — | 1,769 | | [GLM-5.2](/models/zhipu-ai/glm-5-2) | 81.0%* | 46.2%* | 40.6%† | 1,514 | | [Kimi K3](/models/moonshot-ai/kimi-k3) | 88.3% | 67.5% | 28.3%† | 1,686 | | [DeepSeek V4 Pro](/models/deepseek-ai/deepseek-v4-pro) | 87.9% | 62.7% | 25.7%† | 1,307 | *GLM-5.2 Terminal-Bench 2.1/DeepSWE figures are from Zhipu's own published comparison (not yet separately verified as Benchgen leaderboard entries for GLM-5.2). †Agents' Last Exam shown as a percentage here for readability; Benchgen's leaderboard stores it as a 0–1 fraction. Despite being the "Flash" (cost-efficient) tier, GLM-5.3-Flash beats its own larger, more expensive predecessor GLM-5.2 on every metric Zhipu reports, and lands within a few points of Kimi K3 and DeepSeek V4 Pro on coding/agentic benchmarks — while being the only model in this comparison that's both natively multimodal and fully open-weight (MIT). ## Related Models & Benchmarks - Compare: [GLM-5.3](/models/zhipuai/glm-5-3) · [GLM-5.2](/models/zhipu-ai/glm-5-2) · [Kimi K3](/models/moonshot-ai/kimi-k3) · [DeepSeek V4 Pro](/models/deepseek-ai/deepseek-v4-pro) - Benchmarks: [TerminalBench 2.1](/benchmarks/tbench/terminalbench-21) · [DeepSWE](/benchmarks/datacurve/deepswe) · [Agents' Last Exam](/benchmarks/agents-last-exam) · [Humanity's Last Exam](/benchmarks/cais/humanitys-last-exam) · [GDPVal-AA v2](/benchmarks/artificial-analysis/gdpval-aa-v2) ## Frequently Asked Questions

What is GLM-5.3-Flash?

Zhipu AI's (Z.ai) first natively multimodal model in the GLM-5 series — a 320B-total/18B-active MoE released under the MIT license with a 1M-token context window. It outperforms the larger GLM-5.2 on every reported benchmark at roughly one-tenth the price.

Is GLM-5.3-Flash open source?

Yes — it's released under the MIT license with weights on Hugging Face (`zai-org/GLM-5.3-Flash`), fully self-hostable.

Is GLM-5.3-Flash the same as "Ox Alpha"?

GLM-5.3-Flash briefly circulated pre-launch as an anonymous stealth model under third-party branding before Zhipu's identity and the model's real specs (MIT license, 320B/18B-active, HF card) were confirmed.

How does GLM-5.3-Flash compare to the full GLM-5.3?

GLM-5.3 (the proprietary flagship) scores higher on every shared benchmark (e.g. Terminal-Bench 2.1 88.2% vs. 84.3%), but GLM-5.3-Flash is open-weight (MIT) and multimodal, while the flagship remains text-only and API-only pending its own open-weight release.

--- *Specs from Zhipu AI's official HF model card (huggingface.co/zai-org/GLM-5.3-Flash) and Benchgen research. Last updated 2026-08-27.*

GLM-5.3-Flash — Training jobs

Training jobs