Zhipu AI (Z.ai)open-weight
Released: 2026-08-26Z.ai: GLM 5.3 Flash
GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...
Intelligence Index
77.0/ 100
Calibrated multi-domain compositeThroughput Speed
130tok/sec
Output streaming throughputTime To First Token
280ms
Initial chunk server latencyLiveBench Contam-Free
67.8%
September 2026 suite scoreLiveBench Multi-Domain Evaluation Breakdown
Monthly refreshed contamination-free test setsLogical Reasoning70.8%
Coding & Repo Repair69.3%
Mathematics (AIME)68.5%
Data Analysis & Tables65.5%
Language & Comprehension67.0%
Instruction Following67.8%
Sourced Benchmark Evaluations (61)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| Soohak | Active | 57.2% | 98 | 2026-09-12 | independent | Source ↗ |
| Humanity's Last Exam | Active | 34.6% | 93 | 2026-09-12 | independent | Source ↗ |
| SWE-Lancer | Deprecated | 52.4% | 20 | 2026-09-12 | independent | Source ↗ |
| StructFlowBench | Nearing Saturation | 60.6% | 75 | 2026-09-12 | independent | Source ↗ |
| OlympiadBench | Active | 61.7% | 80 | 2026-09-12 | independent | Source ↗ |
| FrontierMath | Active | 21.1% | 98 | 2026-09-12 | independent | Source ↗ |
| FrontierCode | Active | 44.7% | 98 | 2026-09-12 | independent | Source ↗ |
| Omni-MATH | Deprecated | 55.7% | 10 | 2026-09-12 | independent | Source ↗ |
| PutnamBench | Active | 39.2% | 90 | 2026-09-12 | independent | Source ↗ |
| ARC-AGI-3 | Active | 51.1% | 93 | 2026-09-12 | independent | Source ↗ |
| ARC-AGI-2 | Nearing Saturation | 61.3% | 75 | 2026-09-12 | independent | Source ↗ |
| AIME | Nearing Saturation | 71.5% | 75 | 2026-09-12 | independent | Source ↗ |
| FireBench | Active | 59.9% | 95 | 2026-09-11 | independent | Source ↗ |
| Terminal-Bench | Active | 53.9% | 80 | 2026-09-11 | independent | Source ↗ |
| MathArena | Active | 60.2% | 99 | 2026-09-11 | independent | Source ↗ |
| OpenAI-MRCR | Active | 68% | 80 | 2026-09-10 | independent | Source ↗ |
| CodeContests | Deprecated | 50.1% | 10 | 2026-09-10 | independent | Source ↗ |
| SimpleQA | Active | 42.1% | 90 | 2026-09-10 | independent | Source ↗ |
| LongBench-Pro | Active | 66.4% | 80 | 2026-09-09 | independent | Source ↗ |
| SWE-bench Pro | Active | 47.7% | 90 | 2026-09-09 | independent | Source ↗ |
| ComplexBench | Nearing Saturation | 58.5% | 80 | 2026-09-09 | independent | Source ↗ |
| MultiPL-E | Deprecated | 70.8% | 10 | 2026-09-08 | independent | Source ↗ |
| ZeroSCROLLS | Nearing Saturation | 64.8% | 78 | 2026-09-08 | independent | Source ↗ |
| AgentIF | Nearing Saturation | 48.5% | 80 | 2026-09-07 | independent | Source ↗ |
| LiveCodeBench | Saturated | 63.1% | 39 | 2026-09-07 | independent | Source ↗ |
| MATH | Saturated | 73.8% | 25 | 2026-09-07 | independent | Source ↗ |
| InfiniteBench | Nearing Saturation | 68.9% | 65 | 2026-09-07 | independent | Source ↗ |
| Mercury | Deprecated | 67.8% | 10 | 2026-09-06 | independent | Source ↗ |
| SCROLLS | Nearing Saturation | 64.8% | 65 | 2026-09-06 | independent | Source ↗ |
| MMLU-Pro | Nearing Saturation | 74.5% | 75 | 2026-09-06 | independent | Source ↗ |
| RULER | Nearing Saturation | 74.5% | 84 | 2026-09-05 | independent | Source ↗ |
| TabMWP | Saturated | 69.3% | 25 | 2026-09-05 | independent | Source ↗ |
| CyberSecEval | Deprecated | 69.3% | 10 | 2026-09-05 | independent | Source ↗ |
| SWE-bench Verified | Saturated | 60.1% | 25 | 2026-09-04 | independent | Source ↗ |
| MMLU-Redux | Saturated | 79.4% | 25 | 2026-09-04 | independent | Source ↗ |
| QuALITY | Nearing Saturation | 71.3% | 78 | 2026-09-04 | independent | Source ↗ |
| LongBench | Nearing Saturation | 69.7% | 65 | 2026-09-03 | independent | Source ↗ |
| SWE-bench | Deprecated | 52.4% | 10 | 2026-09-03 | independent | Source ↗ |
| MGSM | Nearing Saturation | 71.5% | 65 | 2026-09-03 | independent | Source ↗ |
| AGIEval | Saturated | 67.6% | 25 | 2026-09-02 | independent | Source ↗ |
| QASPER | Nearing Saturation | 67.2% | 65 | 2026-09-02 | independent | Source ↗ |
| TruthfulQA | Saturated | 71.3% | 25 | 2026-09-02 | independent | Source ↗ |
| GSM-Hard | Deprecated | 72.3% | 10 | 2026-09-01 | independent | Source ↗ |
| SuperGLUE | Saturated | 80.2% | 25 | 2026-09-01 | independent | Source ↗ |
| CodeXGLUE | Deprecated | 72.4% | 23 | 2026-08-31 | independent | Source ↗ |
| MT-Bench | Nearing Saturation | 7.3% | 75 | 2026-08-31 | independent | Source ↗ |
| BIG-Bench Hard | Saturated | 77% | 25 | 2026-08-31 | independent | Source ↗ |
| WinoGrande | Saturated | 81.7% | 28 | 2026-08-30 | independent | Source ↗ |
| MBPP+ | Nearing Saturation | 78.5% | 65 | 2026-08-30 | independent | Source ↗ |
| IFEval | Deprecated | 69.9% | 25 | 2026-08-30 | independent | Source ↗ |
| GPQA | Nearing Saturation | 71.3% | 78 | 2026-08-30 | independent | Source ↗ |
| Needle-in-a-Haystack | Saturated | 96.2% | 44 | 2026-08-29 | independent | Source ↗ |
| HumanEval+ | Saturated | 83.2% | 25 | 2026-08-29 | independent | Source ↗ |
| CommonsenseQA | Saturated | 82.5% | 28 | 2026-08-29 | independent | Source ↗ |
| MBPP | Saturated | 81.6% | 25 | 2026-08-28 | independent | Source ↗ |
| GSM8K | Saturated | 81.3% | 25 | 2026-08-28 | independent | Source ↗ |
| GLUE | Saturated | 84.9% | 28 | 2026-08-28 | independent | Source ↗ |
| MMLU | Saturated | 82.6% | 25 | 2026-08-27 | independent | Source ↗ |
| HumanEval | Saturated | 87.8% | 25 | 2026-08-27 | independent | Source ↗ |
| ARC-AGI | Saturated | 73.9% | 25 | 2026-08-27 | independent | Source ↗ |
| LiveBench (Continuous) | Active | 67.8% | 94 | 2025-02-01 | independent | Source ↗ |
Local Hardware Execution (Ollama)Hardware Compatible
# 1. Pull and execute model locally with Ollama
ollama run z-ai-glm-5-3-flash
# 2. Or invoke via local OpenAI-compatible endpoint
curl http://localhost:11434/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "z-ai-glm-5-3-flash",
"messages": [{"role": "user", "content": "Analyze reasoning chains on AIME 2026"}]
}'Evaluation & Sourcing Notes
Scores listed for Z.ai: GLM 5.3 Flash represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.