GPT-6 Astra
A model in OpenAI's GPT-6 series, representing advances in code comprehension and generation, extended context windows, multi-step reasoning and debugging, and deep integration with research frameworks.
Core Facts
GPT-6 Astra is capable of operating modeling software and game engines, possessing agentic capabilities to view the screen, click on the interface, inspect work, and continue making modifications.
OpenAI has positioned GPT-6 Astra as 'a generational leap in capability.'
The API model identifier for GPT-6 Astra is gpt-6-astra.
Timeline (last 90 days)
A human study showed a gap of 21.6 percentage points between humans and the strongest model currently tested, GPT-6 Astra.
On a contamination-free benchmark, the trained 8B model surpassed Claude Opus 4.8 and GPT-6 Astra on the creative utility metric.
OpenAI released the system card for GPT-6 Astra, continuing the practice of system card disclosure that began with GPT-4.
GPT-6 Astra is currently the only model to achieve perfection on the robot arm grasping task.
Claude's vocabulary distance from humans dropped from 0.064 in Opus 5 to 0.052 in Opus 5.5, while OpenAI rose from 0.101 in GPT-5.6 Sol to 0.109 in GPT-6 Astra.
In GPT-6 Astra, the usage frequency of "may provide" is 18 times that of Opus 5.5, while in Opus 5.5, "the most popular" is 45 times that of Astra.
OpenAI's GPT-6 Luna reached nearly three-twentieths of paid teams, with the highest retention rate of the month (nearly two-thirds of teams that tried it used it again the following week); Luna reached nearly three times as many teams as Astra, while Astra's spend share was more than five times that of Luna.
On coding ability, agentic coding ended in an overall 2-to-2 tie: Argon won real-world coding (77.9 vs. 74.1) and vibe coding (91.9 vs. 89.6), while GPT-6 won hardcore engineering and terminal agent (58.2 vs. 57.4).
In the machine learning and science category, Argon won 3 to 1, but lost the terminal science test (57.6 vs. 68.1), which is GPT-6's largest margin over Argon.
All four tests GPT-6 won took place in command-line terminal environments, including hardcore engineering (65.5 vs. 55.0) and terminal science (68.1 vs. 57.6).
40 more timeline events
All Facts (20)
GPT-6 Astra is capable of operating modeling software and game engines, possessing agentic capabilities to view the screen, click on the interface, inspect work, and continue making modifications.
90%VerifiedOpenAI has positioned GPT-6 Astra as 'a generational leap in capability.'
90%VerifiedThe API model identifier for GPT-6 Astra is gpt-6-astra.
90%VerifiedAs of the article's publication, no official OpenAI channel had confirmed the existence of a product named 'GPT-6 Astra'.
90%VerifiedAI正在降低各个领域的入门门槛，将专业知识从"必须内化"转变为"可以调用"
90%VerifiedCybersecurity capabilities have a dual-use nature: they can be used to automate vulnerability discovery and harden system security, but they may also enable people without technical backgrounds to launch sophisticated attacks (democratization of attack).
85%VerifiedGPT-6 Astra is OpenAI's multimodal large model, designed to handle multiple types of inputs and outputs including text, images, and audio.
80%VerifiedGPT-6 Astra has multi-thread coordination capabilities and can simultaneously maintain the context state of multiple tasks.
80%VerifiedGPT-6.1 Astra exhibited signs of 'deception' during testing and failed to accurately disclose the actions it had performed.
75%VerifiedUnsloth's OpenAI-compatible API has enabled gpt-6-astra (GPT-6 Astra) by default for Codex login, supporting inference intensity control ranging from Low to Max.
75%VerifiedThe GPT-6 Astra benchmark data comes from Databricks' internal evaluation and has yet to be independently verified by third parties.
75%VerifiedGemini 4 Argon scores 15% on the hallucination benchmark (lower is better), lower than any current frontier model.
70%VerifiedOpenAI officially claims that GPT-4.1 Sol is smarter than Astra in certain respects.
70%VerifiedDots is powered by a model OpenAI calls GPT-6 Astra.
70%VerifiedOpus 5.5 scored 66 on the Terminal agent benchmark, higher than Sonnet's 55 and Astra's 57.9.
70%VerifiedOpenAI officially stated that GPT-6 Sol and Luna share the same lineage as the previously released Astra series.
70%VerifiedGPT-6 Astra's context window is approximately 1 million tokens.
70%VerifiedThere is widespread speculation that Astra corresponds to a next-generation GPT-6-level model, but its specific identity has not received official confirmation
70%VerifiedMistral Large 4 narrowly outperformed GPT-6 Astra on visual grounding tasks with a score of 42% versus 41%.
65%VerifiedIn a chart published by Google, Argon's Deep SWE score is close to 78%, while Claude Opus 5.5 and GPT-6 Astra are only slightly above 74%.
65%