Contamination MapText / LLM / Model
GPT-6 Astra
- Organisation
- OpenAI
- Family
- GPT · version 6
- Released
- 2026-09-03 (EU AI Act public summary: date of placement on the Union market 3 September 2026)
- Training cutoff
- 2026-04-30
Evidence
“Apr 30, 2026 knowledge cutoff”
developers.openai.com · website · Model page header · retrieved 2026-10-15 “datasets from varying time periods, with some data collected no later than 6 August 2026”
cdn.openai.com · website · Section 1.3 · retrieved 2026-10-15 - Checkpoints
- gpt-6-astra
- Weights
- Closed · proprietary
Training data disclosure: Partial list
Level determined for 0% of benchmark cells.
- TestTest data seen0
- SeenSeen, split unclear0
- TrainTraining data seen0
- Related data seen0
- NoNot seen0
- ?Unknown15
“The training data for GPT-6 Astra includes text from Common Crawl.”
“GPT-6 Astra is not a modification or a fine-tune of a prior model.”
Other declared training data
-
Common Crawl (broad web data: may contain benchmark test sets) · split: unknown
Evidence
“The training data for GPT-6 Astra includes text from Common Crawl.”
cdn.openai.com · website · Section 2.1 · retrieved 2026-10-15 -
Other publicly available datasets (broad, multi-domain text and image datasets from public repositories, online platforms and specialized websites) (broad web data: may contain benchmark test sets) · split: unknown
Evidence
“Other publicly available datasets include broad, multi-domain text and image datasets made available by third parties through public repositories, online platforms, and specialized websites”
cdn.openai.com · website · Section 2.1 · retrieved 2026-10-15 -
Web content crawled by GPTBot (approximately March 2013 - May 2026) (broad web data: may contain benchmark test sets) · split: unknown
Evidence
“GPTBot is used to crawl content that may be used to train OpenAI's generative AI foundation models”
cdn.openai.com · website · Section 2.3 · retrieved 2026-10-15 -
Synthetic data generated with GPT-5.6, GPT-5.5 and GPT-5.4 (not yet matched to a dataset) · split: unknown
Evidence
“OpenAI generated synthetic data using its own general-purpose AI models, including GPT-5.6, GPT-5.5 and GPT-5.4.”
cdn.openai.com · website · Section 2.5 · retrieved 2026-10-15 -
Data from user interactions with ChatGPT and Codex (subject to user settings) (not yet matched to a dataset) · split: unknown
Evidence
“OpenAI may use interactions to train our models, including GPT-6 Astra.”
cdn.openai.com · website · Section 2.4 · retrieved 2026-10-15
Benchmark cells
Classic LLM benchmarks
Unknown MMLU (Measuring Massive Multitask Language Understanding) · training data not disclosed in enough detail
Unknown GSM8K (Grade School Math 8K) · training data not disclosed in enough detail
Unknown MATH (Mathematics Aptitude Test of Heuristics) · training data not disclosed in enough detail
Unknown HumanEval · training data not disclosed in enough detail
Unknown ARC-Challenge (AI2 Reasoning Challenge, Challenge Set) · training data not disclosed in enough detail
Unknown HellaSwag · training data not disclosed in enough detail
Unknown TruthfulQA · training data not disclosed in enough detail
Unknown WinoGrande · training data not disclosed in enough detail
Recent LLM benchmarks
Unknown GPQA Diamond · training data not disclosed in enough detail
Unknown MMLU-Pro · training data not disclosed in enough detail
Unknown LiveCodeBench (code generation, release_v6) · training data not disclosed in enough detail
Unknown SWE-bench Verified · training data not disclosed in enough detail
Unknown AIME 2024 · training data not disclosed in enough detail
Unknown AIME 2025 · training data not disclosed in enough detail
Unknown MATH-500 · training data not disclosed in enough detail
Notes
Knowledge cutoff (docs) is Apr 30, 2026, but the EU summary says some training data was collected as late as 6 August 2026 (crawl period approx. March 2013 - May 2026). No GPT-6 system card was fetched. GPT-6 Luna (knowledge cutoff May 18, 2026), GPT-6 Sol and GPT-6.1 Sol are separate models not covered here.
Sources
- Public Summary of Training Content for GPT-6 Astra (EU AI Act), v1, 2 September 2026 · website · retrieved 2026-10-15
- OpenAI API docs: GPT-6 Astra model page · website · retrieved 2026-10-15
- OpenAI API docs: Models · website · retrieved 2026-10-15