GLM-5.3-Flash Benchmarks, Specifications & Availability
Explore GLM-5.3-Flash from Z.ai: published specifications, source-linked vendor benchmarks, independent evaluator coverage and recorded pricing when available.
Compare GLM-5.3-Flash with other models →Explore data coverage
Published specifications
- Provider
- Z.ai
- Access
- Open weights
- License
- MIT
- Context window
- 1M
- Total parameters
- 320B
- Active parameters
- 18B
- Released
- 2026-08-26
- Modalities
- text, image, video
- Family
- GLM-5.3
Model card · Announcement · Website · OpenRouter
Model notes
First natively multimodal GLM-5 series model; efficiency tier. Open weights verified on Hugging Face https://huggingface.co/zai-org/GLM-5.3-Flash (2026-09-25).
Pricing · OpenRouter
Dated cached OpenRouter rates in USD per 1M tokens. Open the dashboard for live enhancements. Per-metric endpoint minima can refer to different providers; they are not a guaranteed combined rate from one endpoint.
| Tier | Input / 1M | Output / 1M | Cached input / 1M | Cache write / 1M | Date & source |
|---|---|---|---|---|---|
| Default | [object Object] | [object Object] | [object Object] | — | 2026-10-03 · OpenRouter source |
Recorded pricing notes
Re-pulled 2026-09-25 from OR API; was in/out/cache 0.15/0.5/0.05 on 2026-09-23.; min-healthy endpoint minima 2026-10-01; discount flag 0.5 on DeepInfra (undocumented, not applied); min-healthy endpoint minima 2026-10-01; discount flag 0.5 on DeepInfra (undocumented, not applied); min-healthy endpoint minima 2026-10-02; discount flag 0.5 on DeepInfra (undocumented, not applied); min-healthy endpoint minima 2026-10-03; discount flag 0.5 on DeepInfra (undocumented, not applied); discount flag 0.5 on DeepInfra (undocumented, not applied)
Official / vendor benchmarks
Default headline records. Own-vendor, peer-vendor and third-party provenance remain visible in evidence; configurations may differ.
| Benchmark / evaluator | Headline score | Evidence |
|---|---|---|
| AA Intelligence Index (vendor-cited) | 57unknown effort | All 1 recorded result & sources57 · raw 57 index Headline · unknown effort · Own vendor Source/record date: 2026-08-26 AA Intelligence Index v4.1.1 as cited by z.ai ($0.045/task discounted) https://z.ai/blog/glm-5.3-flash |
| Agents' Last Exam | 26.3%unknown effort | All 2 recorded results & sources26.3% · raw 26.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash26.3% · raw 26.3 % Alternative · unknown effort · Peer vendor Source/record date: 2026-09-27 As reported by NaiveAI. https://naive.ai/en/research/ |
| AutomationBench v1.0.6 | 48.8%unknown effort | All 1 recorded result & sources48.8% · raw 48.8 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash |
| DeepSWE v1.1 | 63.4%unknown effort | All 3 recorded results & sources63.4% · raw 63.4 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash63.4% · raw 63.4 % Alternative · unknown effort · Peer vendor Source/record date: 2026-09-27 As reported by NaiveAI. https://naive.ai/en/research/63.4% · raw 63.4 % Alternative · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. | Same value as current headline; kept as corroboration, non-headline. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| ExtractBench Short | 96.3%unknown effort | All 1 recorded result & sources96.3% · raw 96.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 ExtractBench Short from HF eval badge https://huggingface.co/zai-org/GLM-5.3-Flash |
| GDPval-AA v2 | 1773unknown effort | All 1 recorded result & sources1773 · raw 1773 Elo Headline · unknown effort · Own vendor Source/record date: 2026-08-26 Elo GDPval-AA v2 https://z.ai/blog/glm-5.3-flash |
| Humanity's Last Exam (w/ tools) | 55.3%unknown effort | All 1 recorded result & sources55.3% · raw 55.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash |
| MVBench | 77.8%unknown effort | All 1 recorded result & sources77.8% · raw 77.8 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 MVBench https://z.ai/blog/glm-5.3-flash |
| NL2Repo | 56.3%unknown effort | All 1 recorded result & sources56.3% · raw 56.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 NL2Repo on z.ai blog table https://z.ai/blog/glm-5.3-flash |
| NL2Repo-Bench | 56.3%unknown effort | All 1 recorded result & sources56.3% · raw 56.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash |
| OSWorld 2.0 | 59.1%unknown effort | All 1 recorded result & sources59.1% · raw 59.1 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 OSWorld 2.0 https://z.ai/blog/glm-5.3-flash |
| Terminal-Bench 2.1 | 84.3%unknown effort | All 3 recorded results & sources84.3% · raw 84.3 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash84.3% · raw 84.3 % Alternative · unknown effort · Peer vendor Source/record date: 2026-09-27 As reported by NaiveAI. https://naive.ai/en/research/84.3% · raw 84.3 % Alternative · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. | Same value as current headline; kept as corroboration, non-headline. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| Toolathlon-Verified | 78.4%unknown effort | All 1 recorded result & sources78.4% · raw 78.4 % Headline · unknown effort · Own vendor Source/record date: 2026-08-26 https://z.ai/blog/glm-5.3-flash |
| Z.ai Code Bench v1.0 | 29%max effort | All 1 recorded result & sources29% · raw 29 % Headline · max effort · Own vendor Source/record date: 2026-08-26 max effort; vs Opus 4.8 29.5 https://z.ai/blog/glm-5.3-flash |
| OSWorld-Verified | 62.3%unknown effort | All 1 recorded result & sources62.3% · raw 62.3 % Headline · unknown effort · Peer vendor Source/record date: 2026-09-08 As reported in Nex-N2.5-Pro model card comparison table https://huggingface.co/nex-agi/Nex-N2.5-Pro |
| OSWorld-G | 83.3%unknown effort | All 1 recorded result & sources83.3% · raw 83.3 % Headline · unknown effort · Peer vendor Source/record date: 2026-09-08 As reported in Nex-N2.5-Pro model card comparison table https://huggingface.co/nex-agi/Nex-N2.5-Pro |
| SWE-MM | 20.6%unknown effort | All 1 recorded result & sources20.6% · raw 20.6 % Headline · unknown effort · Peer vendor Source/record date: 2026-09-08 As reported in Nex-N2.5-Pro model card comparison table https://huggingface.co/nex-agi/Nex-N2.5-Pro |
| FrontierCode 1.1 Main | 31.8%unknown effort | All 1 recorded result & sources31.8% · raw 31.8 % Headline · unknown effort · Own vendor Source/record date: 2026-09-02 Cognition FrontierCode 1.1 Main weighted score; model added Sep 2, 2026 per Cognition changelog; score 31.8% on board export mirrored by evals.report (sourceUrl cognition.com, verifiedStatus official, snapshot 2026-09-07) https://cognition.com/frontiercode |
| SWE-Pro | 63.1%unknown effort | All 1 recorded result & sources63.1% · raw 63.06 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| Terminal-Bench 4.0 | 32.8%unknown effort | All 1 recorded result & sources32.8% · raw 32.8 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| HealthBench Professional | 49.1%unknown effort | All 1 recorded result & sources49.1% · raw 49.07 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| DRACO | 78.6%unknown effort | All 1 recorded result & sources78.6% · raw 78.55 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| WideSearch | 80.2%unknown effort | All 1 recorded result & sources80.2% · raw 80.24 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| Humanity's Last Exam | 39.9%unknown effort | All 1 recorded result & sources39.9% · raw 39.9 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
| MultiChallenge | 62.6%unknown effort | All 1 recorded result & sources62.6% · raw 62.6 % Headline · unknown effort · Third-party Source/record date: 2026-09-30 Via ThreatFrontier transcription of Ant launch table; original pixels not yet re-read. Effort undisclosed. https://threatfrontier.com/articles/ling-3-1-flash-ant-group-best-flash-model-yet-cybergym |
Independent evaluators
Evaluator harnesses are distinct from vendor measurements. Missing coverage is not a failed test.
| Benchmark / evaluator | Headline score | Evidence |
|---|---|---|
| Intelligence Index · Artificial Analysis | 42unknown effort | All 1 recorded result & sources42 · raw 42 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-03 Current AA model leaderboard; https://artificialanalysis.ai/models/glm-5-3-flash |
| Cost per Intelligence Index task · Artificial Analysis | $0.25unknown effort | All 1 recorded result & sources$0.25 · raw 0.25 USD Headline · unknown effort · Independent evaluator Source/record date: 2026-10-03 Current AA model leaderboard; https://artificialanalysis.ai/models/glm-5-3-flash |
| Output speed · Artificial Analysis | 54 tok/sunknown effort | All 1 recorded result & sources54 tok/s · raw 54 tok/s Headline · unknown effort · Independent evaluator Source/record date: 2026-10-03 Current AA model leaderboard; Median output tokens/s; leaderboard rounds to whole tokens. https://artificialanalysis.ai/models/glm-5-3-flash |
| Vals Index · Vals AI | 47.2%unknown effort | All 1 recorded result & sources47.2% · raw 47.22 % Headline · unknown effort · Independent evaluator Source/record date: 2026-09-23 https://www.vals.ai/benchmarks/vals_index |
| Bugs fixed /105 · Bug Hunt Bench | 17.7 fixesmax effort | All 4 recorded results & sources17.7 fixes · raw 17.7 fixes Headline · max effort · Independent evaluator Source/record date: 2026-10-03 Harness: Claude Code / Z.ai API; effort max; 3 runs; evaluation 2026-09-15. Best documented score for this effort in Oct 1 README. Headline: best documented model run. https://github.com/phuryn/bug-hunt-bench16 fixes · raw 16 fixes Alternative · high effort · Independent evaluator Source/record date: 2026-10-03 Harness: Claude Code / Z.ai API; effort high; 1 runs; evaluation 2026-09-13. Best documented score for this effort in Oct 1 README. https://github.com/phuryn/bug-hunt-bench13 fixes · raw 13 fixes Alternative · unknown effort · Independent evaluator Source/record date: 2026-10-03 Harness: Claude Code / OpenRouter; effort default; 1 runs; evaluation 2026-08-27. Best documented score for this effort in Oct 1 README. https://github.com/phuryn/bug-hunt-bench9 fixes · raw 9 fixes Alternative · low effort · Independent evaluator Source/record date: 2026-10-03 Harness: Claude Code / Z.ai API; effort low; 1 runs; evaluation 2026-09-13. Best documented score for this effort in Oct 1 README. https://github.com/phuryn/bug-hunt-bench |
| GDPval-AA Elo · Artificial Analysis | 1641unknown effort | All 1 recorded result & sources1641 · raw 1641 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-09-23 GDPval-AA v2.1 Elo; GLM 5.3 Flash https://artificialanalysis.ai/evaluations/gdpval-aa |
| Hard board % of roofline · KernelBench (community board) | 6.5%unknown effort | All 1 recorded result & sources6.5% · raw 6.5 % Headline · unknown effort · Independent evaluator Source/record date: 2026-09-23 Community board lists model as GLM-5.3 Flash (slug ox-alpha). Hard deck % of roofline (3/6); Mega 13.64×; CUDA 4.2% 2/4. Not KernelGen 1P. https://kernelbench.com/models/ox-alpha |
| Vibe Code Bench v1.1 · Vals AI | 30.8%unknown effort | All 1 recorded result & sources30.8% · raw 30.76 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-03 Refreshed from current Vals leaderboard. Harness: OpenHands. Cost/test $2.35. https://www.vals.ai/benchmarks/vibe-code |
| ARC-AGI-1 · ARC Prize | 91%max effort | All 3 recorded results & sources91% · raw 91 % Headline · max effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties. https://arcprize.org/results/zai-glm-5-3-flash71.8% · raw 71.8 % Alternative · high effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. https://arcprize.org/results/zai-glm-5-3-flash47% · raw 47 % Alternative · low effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. https://arcprize.org/results/zai-glm-5-3-flash |
| ARC-AGI-2 · ARC Prize | 65.8%max effort | All 3 recorded results & sources65.8% · raw 65.8 % Headline · max effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. Headline: best verified value; source matrix effort order breaks ties. https://arcprize.org/results/zai-glm-5-3-flash50.1% · raw 50.1 % Alternative · high effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. https://arcprize.org/results/zai-glm-5-3-flash27.9% · raw 27.9 % Alternative · low effort · Independent evaluator Source/record date: 2026-10-03 Semi-Private verified matrix. Published verified configuration. https://arcprize.org/results/zai-glm-5-3-flash |
Read how we select and source scores or the comparison guide.