AI lab · 16 models tracked
Z.ai models
Every Z.ai model released in the last 18 months, with release dates, context windows, API prices and benchmark scores. Newest: GLM 5.3 Prime, 26d ago.
- Models · 30 days
- 1 of 16
- Latest release
- 26d ago
- Cheapest input / 1M
- $0.047 GLM 5.3 Flash
- Largest context
- 1M GLM 5.3 FlashX
Release history
Timeline
Z.ai ships a new model about every 28 days. Each dot is a release; click one to open it.
Lineup
All Z.ai models
| Model | Released | Context | Input / 1M | Output / 1M | GPQA Diamond |
|---|---|---|---|---|---|
GLM 5.3 Primez-ai/glm-5.3-prime |
Sep 23, 2026 | 1M131K out | $2.80 | $8.80 | — |
GLM 5.3 FlashXz-ai/glm-5.3-flashx |
Sep 18, 2026 | 1M131K out | $0.37 | $1.25 | — |
GLM 5.3 Flashz-ai/glm-5.3-flash |
Aug 26, 2026 | 1M944K out | $0.047 | $0.16 | 90.1% |
GLM 5.3z-ai/glm-5.3 |
Aug 18, 2026 | 1M944K out | $0.079 | $1.40 | 90.9% |
GLM 5.2z-ai/glm-5.2 |
Jun 16, 2026 | 1M944K out | $0.62 | $2.42 | 91.9% |
GLM 5.1z-ai/glm-5.1 |
Apr 7, 2026 | 205K128K out | $0.97 | $3.04 | 89.9% |
GLM 5V Turboz-ai/glm-5v-turbo |
Apr 1, 2026 | 203K131K out | $1.20 | $4.00 | — |
GLM 5 Turboz-ai/glm-5-turbo |
Mar 15, 2026 | 203K131K out | $1.20 | $4.00 | — |
GLM 5z-ai/glm-5 |
Feb 11, 2026 | 205K128K out | $0.60 | $1.92 | 87.8% |
GLM 4.7 Flashz-ai/glm-4.7-flash |
Jan 19, 2026 | 200K118K out | $0.061 | $0.40 | 60.5% |
GLM 4.7z-ai/glm-4.7 |
Dec 22, 2025 | 205K131K out | $0.54 | $1.98 | 83.3% |
GLM 4.6Vz-ai/glm-4.6v |
Dec 8, 2025 | 131K33K out | $0.30 | $0.90 | — |
GLM 4.6z-ai/glm-4.6 |
Sep 30, 2025 | 205K131K out | $0.55 | $2.20 | — |
GLM 4.5Vz-ai/glm-4.5v |
Aug 11, 2025 | 66K16K out | $0.60 | $1.80 | — |
GLM 4.5z-ai/glm-4.5 |
Jul 25, 2025 | 131K98K out | $0.60 | $2.20 | — |
GLM 4.5 Airz-ai/glm-4.5-air |
Jul 25, 2025 | 131K98K out | $0.13 | $0.85 | — |
Benchmarks
Z.ai's best scores
The top Z.ai model on each benchmark, and where it ranks among all 222 tracked models.
| Benchmark | Best Z.ai model | Score | Overall rank |
|---|---|---|---|
| FrontierSWESoftware engineering | GLM 5.3 | 30.2% | #10of 19 |
| DeepSWESoftware engineering | GLM 5.3 | 69.0% | #7of 26 |
| CursorBenchSoftware engineering | GLM 5.3 | 42.6% | #8of 16 |
| FrontierCodeSoftware engineering | GLM 5.3 | 40.1% | #25of 41 |
| SWE-bench VerifiedSoftware engineering | GLM 5.2 | 78.7% | #4of 26 |
| Terminal-BenchSoftware engineering | GLM 5 | 52.4% | #16of 40 |
| GBAEvalSoftware engineering | GLM 5.2 | 0.0% | #21of 23 |
| GSOSoftware engineering | GLM 4.5 Air | 2.9% | #20of 20 |
| SciCodeSoftware engineering | GLM 5.3 | 59.0% | #10of 93 |
| APEX-AgentsAgents & computer use | GLM 5.3 | 56.6% | #15of 51 |
| Vending-Bench 2Agents & computer use | GLM 5.2 | $8,314 | #11of 61 |
| PostTrainBenchAgents & computer use | GLM 5.2 | 31.7% | #6of 11 |
| Lech Mazur WritingWriting & content | GLM 4.5 | 7.34/10 | #12of 13 |
| GDP.pdfWriting & content | GLM 5.3 Flash | 14.0% | #31of 38 |
| SimpleQA VerifiedWriting & content | GLM 5.3 | 41.0% | #41of 66 |
| WebDev ArenaDesign & visual | GLM 5.3 | 1614 | #17of 88 |
| ExploitBenchCybersecurity | GLM 5.1 | 18.1% | #6of 8 |
| GPQA DiamondReasoning & knowledge | GLM 5.2 | 91.9% | #21of 92 |
| Humanity's Last ExamReasoning & knowledge | GLM 4.5 | 8.3% | #22of 23 |
| ARC-AGI-2Reasoning & knowledge | GLM 5.3 Flash | 65.8% | #25of 63 |
| ARC-AGI-1Reasoning & knowledge | GLM 5.3 Flash | 91.0% | #26of 63 |
| SimpleBenchReasoning & knowledge | GLM 5.2 | 58.8% | #24of 43 |
| CritPtReasoning & knowledge | GLM 5.2 | 20.9% | #20of 96 |
| Chess PuzzlesReasoning & knowledge | GLM 5.3 | 37.0% | #38of 70 |
| LMCAReasoning & knowledge | GLM 5.3 | 55.5% | #13of 96 |
| FrontierMathMaths | GLM 5.3 | 68.8% | #26of 72 |
| FrontierMath Tier 4Maths | GLM 5.3 | 29.3% | #29of 62 |
| ProofBenchMaths | GLM 5.3 | 49.0% | #35of 70 |
| OTIS Mock AIMEMaths | GLM 5.3 Flash | 93.9% | #43of 90 |
Directory