AA-Briefcase
1.1 / combined Elo ↗- GLM-5.3-Flash:
- GPT-6 Astra:
Compare models
Find the right model, side by side.
Updated: Data as of:
First model
Second model
Model specifications
| Specifications |  GLM-5.3-FlashZ.ai |  GPT-6 AstraOpenAI |
|---|---|---|
| Provider | ||
| Context sizetokens | 1,000,000 | 1,000,000 |
| Cutoff date | Not reported | 2026-04-30 |
| Input costUSD / 1M tokens | $0.15 | $10.00 |
| Output costUSD / 1M tokens | $0.50 | $50.00 |
| Max outputtokens | Not reported | 128,000 |
| LatencyTTFT · seconds | 1.546 | 365.127 |
| Speedtokens / sec | 75.174 | 70.564 |
Benchmark performance
14 shared benchmark comparisons. Versions and setups matched.
GLM-5.3-FlashGPT-6 Astra
AutomationBench-AA
1.0.6 / AA private 657 ↗Tools enabled- GLM-5.3-Flash:
- GPT-6 Astra:
- GLM-5.3-Flash:
- GPT-6 Astra:
GDPval-AA
2.1 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
GPQA Diamond
Diamond / Independent evaluation ↗- GLM-5.3-Flash:
- GPT-6 Astra:
HLE-Diamond
1,000 questions / no tools ↗No tools- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Agentic
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Browser Use
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Coding
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Data Analysis
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Design
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Reasoning
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
LiveBench: Research
2.0 / 2026-10-25 ↗- GLM-5.3-Flash:
- GPT-6 Astra:
Terminal-Bench
4.0 / mini-swe-agent ↗mini-swe-agent- GLM-5.3-Flash:
- GPT-6 Astra:
