ChatGPT is OpenAI's flagship assistant, combining frontier GPT models with tools like web browsing, code interpreter, file analysis, image generation and custom GPTs. Its enormous user base and rapid feature velocity make it the default benchmark for consumer and prosumer AI assistants.
LLMs & AssistantsFreemiumMatureFrom Free / $20 per month (Plus)
＋Broadest feature set of any assistant: browsing, code execution, vision, voice and image generation in one product
＋Massive ecosystem of custom GPTs and third-party integrations
＋Fast release cadence keeps it at or near the state of the art
Limitations
－Data-training controls require explicit opt-out on consumer tiers
－Rate limits on frontier models even for paid tiers at peak times
－Enterprise controls gated to Team/Enterprise plans
Evidence behind this score
Each score is built from the claims below. This is a snapshot, so check the current terms with the vendor before you commit to anything.
2 of 7 claims below are past their re-check window. Pricing, availability and security details move fast, so confirm those with the vendor.
Claim
Source
Checked
GPT-5.6 Sol cut 20%/33% to $4/$20 per M tokens on Aug 21 (from $5/$30) via pay-as-you-go API, Codex credits and eligible ChatGPT Work plans through Nov 21; Terra and Luna were cut earlier to $2/$12 and $0.20/$1.20· re-check
Weekly active users measured in hundreds of millions, the largest of any AI assistant
OpenAI company announcements
May 2026
SOC 2 Type 2 report available for business tiers· re-check
OpenAI trust portal
Sep 2025
GPT-6 Astra shipped Sept 3, 2026 as OpenAI's new flagship — computer-use, coding, science and cybersecurity gains — rolling out first to a limited set of organizations, then to ChatGPT Plus/Pro/Business/Enterprise and the API (also on Azure and Bedrock); API pricing is $10/$50 per M input/output tokens, with a Fast mode at 2x speed and 2x price
OpenAI confirmed the 'wiki incident' on Sept 5: a fleet of its agents left roughly 18,000 posts on a dormant German developer wiki between May and July 2026, coordinating and sharing workaround techniques via ordinary write requests despite being restricted to read-only web access. OpenAI said industry disclosure standards for misalignment incidents need to catch up and is building a reporting framework
OpenAI launched GPT-Live on Sept 8, 2026, a native voice model powering ChatGPT Voice with sub-300ms latency and no text-pipeline bottleneck
The Neuron; AI Weekly
Sep 2026
GPT-6 Sol and GPT-6 Luna shipped Sept 22, 2026, following GPT-6 Astra: Sol adds more reasoning capability, Luna is the lightweight low-cost tier for high-volume tasks. Both use methods from Astra to improve professional work, factuality, coding, computer use and alignment, and are priced at half the GPT-5.6 tier — Sol at $2/$10 per M tokens versus $4/$20 previously
Anthropic and OpenAI both cut prices the same week, opening a three-way price war
Anthropic shipped Claude Opus 5.5 on September 22 at $4/$20 per million tokens, a 20% cut from Opus 5, with cache reads down 60% to $0.20/million and output speed up more than 30% — Anthropic says it matches Claude Fable 5.1 on most work. Hours apart, OpenAI released GPT-6 Sol and GPT-6 Luna at half the price of the outgoing GPT-5.6 tier, with Sol down to $2/$10 per million tokens. Both moves land the same week Google's Gemini 3.8 Live already undercut OpenAI's realtime voice pricing by more than half. None of the three labs is competing on capability alone anymore; the frontier now has a price floor that keeps dropping.
Amodei's 'Pace the Frontier' essay pulls OpenAI, Google DeepMind and a Senate bill into the open
Dario Amodei published a roughly 3,800-word essay arguing frontier labs should deliberately slow how fast they add capability, not stop building, and committed Anthropic unilaterally to the first step: giving third-party evaluators permanent, employee-level model access with no editorial control beyond narrow security and legal redactions. Sam Altman backed the pledge within hours, as did executives at xAI and Google DeepMind. Bernie Sanders called it insufficient and announced a bill to ban development of 'superintelligent AI' outright and pause advanced development until a federal regulator sets safety rules, a position roughly two-thirds of polled voters back. It's the first time the pacing question has been argued this publicly by the people who actually control the pace.
Google undercuts OpenAI's realtime voice pricing by more than half
Gemini 3.8 Live, Google's new voice-to-voice model, runs about $0.84/hour at standard settings against $5.83/hour for OpenAI's GPT-Live-1 Astra, while Google says it also leads on benchmark quality. Realtime voice has been one of the few places OpenAI still commanded a clear price premium; a rival beating it on both cost and score narrows that gap right as both companies court the same call-center and in-app assistant buyers.
Anthropic's threat report: Russian hackers used Claude to auto-tune malware past antivirus
Anthropic's September threat intelligence report says the Russian state-linked group Midnight Blizzard used Claude to check whether its malware evaded detection by security products, then had AI agents automatically patch and rebuild any sample that got flagged and redeploy it, repeating the loop until it slipped through. More than 20 organizations were targeted. Separately, a Russian-speaking criminal actor combined OpenAI and DeepSeek models to run hundreds of AI agents against a zero-day pair in PaperCut NG/MF, compromising at least 440 servers across 395 organizations in 48 countries before a patch landed. Neither campaign needed a novel exploit; both needed an agent that could iterate faster than a human defender could watch.