OpenAI
GPT-6 Astra
GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.
openai/gpt-6-astra
Context
1.1M
Output
128K
Released
2026-09-04
Knowledge cutoff
2026-04-30
Types
Reasoning
Tool use
Structured output
Attachments
Open weights
Provider pricing
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $10.00 / $50.00
- Context
- 1.1M
- Status
- Stable
- Price
- $13.00 / $65.00
- Context
- 1.1M
- Status
- Stable
Benchmarks
Agents' Last Exam
59.3
OSWorld
72.6 · partial score
ScreenSpot-Pro
92.7 · accuracy
AutomationBench
41.4 · success rate
BenchCAD
95.9 · geometric overlap
BrowseComp
91.5 · accuracy
Terminal-Bench
57.9 · success rate
DeepSWE
74.1 · resolve rate
FrontierCode
64.5 · score
Terminal-Bench Science
64.6 · success rate
FrontierMath
97.6 · accuracy
GPQA Diamond
96 · accuracy
Humanity's Last Exam
57.2 · accuracy
ExploitBench
100 · success rate
SRE-Bench
88 · success rate
ARC-AGI-3
99.9 · RHAE
Artificial Analysis Intelligence Index
61 · index score
Artificial Analysis Coding Agent Index
67 · index score