Summary
OpenAI released GPT-5 on August 7, 2025, introducing a unified system that routed between a fast model and a deeper reasoning model. The launch brought GPT-5 to ChatGPT, including the free tier, and introduced GPT-5 pro for subscribers to the Pro plan.
Editorial correction, October 26, 2026
The primary announcement is dated August 7, 2025; April 14, 2025 was the GPT-4.1 API launch. The original April date is corrected in the event metadata. The original body below retains the mistaken date and single-model description; the announcement describes a routed system, rather than a completed merger into one model.
Revision record with the prior summary and metadata.
The original narrative below is preserved as published; use the revised summary and this note for the corrected account.
What Happened
On April 14, 2025, OpenAI released GPT-5, positioning it as the successor to both GPT-4o and the o-series reasoning models. Rather than maintaining separate model families for standard and reasoning tasks, GPT-5 integrated chain-of-thought reasoning as a built-in capability that could be engaged as needed — similar to the approach Anthropic had taken with Claude 3.7 Sonnet's extended thinking.
GPT-5 demonstrated broad improvements across the board: stronger multilingual performance, reduced hallucination, improved instruction following, and competitive reasoning scores. It was made available across ChatGPT tiers (including free users for basic access) and through the API.
The model was positioned as OpenAI's new foundation, replacing the fragmented lineup of GPT-4o, GPT-4o mini, o1, o3, and various specialized models with a more streamlined offering.
Why It Matters
GPT-5's release was significant both for what it contained and what it signaled. The unification of standard and reasoning capabilities confirmed that test-time reasoning was becoming a standard feature rather than a specialized model type — validating the approach that both Anthropic and DeepSeek had already taken.
The model also represented OpenAI's answer to increasing competitive pressure. By early 2025, the landscape had changed dramatically from the GPT-4 era: Anthropic, Google, and DeepSeek all offered frontier-competitive models, and OpenAI's once-dominant position had eroded. GPT-5 was an attempt to reassert leadership through a comprehensive model that excelled across multiple dimensions simultaneously.
Whether GPT-5 represented a genuine leap forward or an incremental improvement over GPT-4o + o3 remained debated, reflecting the broader question of whether the era of dramatic, generation-defining capability jumps was giving way to more gradual, multi-axis improvement.
§ How to read the metadata
- Landmark
- Fundamentally alters the trajectory; 2–5 per year.
- Major
- Meaningfully shifts the landscape; 2–4 per month.
- Notable
- Worth documenting; significance can be upgraded later.
- Confidence
- High = primary sources corroborate. Medium = credible secondary only. Low = provisional. Disputed = credible sources disagree.
- Contestation
- Uncontested = no formal challenge. Contested = at least one challenge open. Superseded = replaced by a later entry. Unresolved = dispute still open.