OpenAI Launches GPT-6 Astra, Matches Claude Fable Pricing at $10/$50 per Million Tokens
OpenAI has begun rolling out GPT-6 Astra, priced at $10/million input and $50/million output tokens to match Claude Fable. The model claims a 99.9% score on ARC-AGI 3 using a custom harness and leads on security and long-context benchmarks, though it trails Fable on general intelligence rankings.
OpenAI has started rolling out GPT-6 Astra, its newest flagship model, to a limited set of organizations today, with broader availability for ChatGPT Plus, Pro, Business, and Enterprise users—plus API and AWS access—expected over the coming days. The API model identifier will be gpt-6-astra.
Astra is priced at $10 per million input tokens and $50 per million output tokens, identical to Anthropic's Claude Fable 5 and 5.1. The matching price point signals OpenAI is positioning Astra as a direct competitor to Fable in the high-end model tier.
Benchmark Claims
According to OpenAI, Astra scores 99.9% on the ARC-AGI 3 benchmark, a test released in March. However, the ARC-AGI blog notes this score was achieved using a custom "Provider Adapter harness" at a cost of $19,000, which "preserves opaque reasoning state between requests and uses compaction for longer conversations, allowing the model to reuse prior work." Using the default ARC-AGI harness, Astra scored 62.7% at a cost of $26,000. Claude Fable 5 does not yet have a published ARC-AGI 3 result for comparison.
On security-related tasks, OpenAI claims Astra scores 100% on ExploitBench, up from GPT-5.6 Sol's 78.5%. On ExploitGym, Astra scored 42.4% versus Sol's 30.3%. On SRE-Bench binary reverse engineering, Astra hit 99.2% within four attempts compared to Sol's 68.7%. These security gains follow OpenAI's recent Hugging Face security incident.
Long-context performance also improved: on OpenAI's internal eight-needle benchmark, Astra scored 100% at the 256K–512K token range and 96.3% at 512K–1M tokens.
Third-Party Results Are Mixed
Independent benchmarking firm Artificial Analysis reports a more complicated picture. On its Intelligence Index, Astra scores 61—tied with GPT-5.6 Sol but five points below Claude Fable 5.1 (max with fallback configuration). The firm also notes Astra trails Meta's newly released Muse Spark 1.3 (max) on the same index.
Astra performs better on Artificial Analysis's Coding Agent Index, where it leads the cost-efficiency frontier. At max effort settings, Astra costs roughly the same as GPT-5.6 Sol while scoring two points higher, and delivers the same score as Claude Fable 5 at less than half the cost per task.
What This Means
The gap between OpenAI's self-reported ARC-AGI 3 score and the underlying methodology is the story here. A 99.9% result achieved through a custom harness that preserves reasoning state across requests is not directly comparable to the 62.7% scored under standard test conditions—and readers should treat the headline number with corresponding skepticism until independent labs replicate results under consistent harnesses.
The broader competitive picture: Astra does not clearly beat Claude Fable on general intelligence per Artificial Analysis's independent index, despite matching its price. Where Astra distinguishes itself is coding-agent cost efficiency and security-specific benchmarks—areas directly relevant to enterprise and DevSecOps buyers. Whether that's enough to justify switching from Fable will depend on real-world testing rather than benchmark claims from either lab. As is often the case with same-day launch coverage, independent verification remains pending.
Related Articles
OpenAI Releases GPT-6 Astra, First Model to Cross 'Critical' Cybersecurity Threshold
OpenAI has begun rolling out GPT-6 Astra, the first model to reach the company's internal 'Critical' cybersecurity threshold. Access is being phased, with companies in OpenAI's Daybreak cybersecurity program getting priority following added safeguards after a prior model containment breach.
OpenAI Launches GPT-6 Astra, Claims State-of-the-Art Computer Use and 98% on FrontierMath Tier 4
OpenAI has launched GPT-6 Astra, claiming state-of-the-art results on computer use, coding, and scientific reasoning benchmarks. The model is rolling out to a limited set of organizations first, with general ChatGPT availability expected within days.
OpenAI Launches GPT-6 Astra, Says the Model May Already Qualify as AGI
OpenAI has released GPT-6 Astra, its most capable model yet, with benchmark scores the company says surpass GPT-5.6 Sol and Anthropic's Fable 5 models. President Greg Brockman called it a step into the 'AGI era,' though OpenAI acknowledges there's no agreed-upon threshold for that term.
OpenAI's GPT-6 Astra Reportedly Automates AI Engineering Tasks at Under $6 an Hour, According to Latent Space Testing
A Latent Space report describes GPT-6 Astra, a new OpenAI model the blog says can autonomously handle AI engineering tasks—training models, labeling data, deploying systems—at an estimated cost of under $6 per hour. The claims, including 97.6% on FrontierMath and 99.9% on ARC-AGI-3, come from independent blog testing rather than an official OpenAI announcement.
Comments
Loading...