OpenAI launches GPT‑5; aggressive pricing signals potential LLM price war

OpenAI’s new flagship raises the bar on capability—and lowers the bar on cost

OpenAI released GPT‑5 on August 7, 2025, positioning it as a unified system that routes between fast responses and deeper “thinking” when tasks require it. ChatGPT now defaults to GPT‑5 for signed‑in users, with Plus/Pro tiers getting higher limits, and a Pro variant for extended reasoning. TechCrunch reports the rollout to free users and the shift toward more agentic behavior that can execute tasks like building apps or drafting research briefs. TechCrunch; OpenAI.

See related coverage on this site: GPT-5 tag and Model Releases.

What’s new under the hood

OpenAI highlights advances in coding, instruction following, tool use and factuality. On developer benchmarks, GPT‑5 scores 74.9% on SWE‑bench Verified and introduces new API controls like a verbosity setting and minimal reasoning mode, plus support for “custom tools.” These details underscore a push toward robust agent workflows in production. OpenAI.

From a user‑experience angle, GPT‑5 simplifies model selection by auto‑deciding when to think longer, while adding new ChatGPT features and personalities. The strategic aim: make high‑end reasoning feel default and accessible rather than gated behind model pickers. TechCrunch.

Pricing that could reshape the market

OpenAI priced GPT‑5’s API aggressively at $1.25 per million input tokens and $10 per million output tokens (with smaller, cheaper sizes also available). That undercuts many frontier rivals on headline rates and may force responses from Anthropic, Google, and others. OpenAI; TechCrunch.

TechCrunch frames the move as potentially catalyzing a long‑anticipated LLM price war, noting GPT‑5’s competitiveness versus peer models and the likelihood that lower costs could ripple through the stack for AI-native startups and enterprise adopters. TechCrunch.

Industry context and impact

Coverage from TechXplore (AFP/AP) situates the launch in a broader race, with OpenAI touting GPT‑5 as a step toward more capable agentic systems while acknowledging that AGI remains a longer‑term goal. The publication also notes the product’s emphasis on coding and real‑world task completion—a practical shift from chat to execution. TechXplore; TechXplore.

For enterprises, the combination of stronger reasoning, more faithful tool use, and lower token prices could tilt build‑vs‑buy decisions toward GPT‑5‑powered apps and agents, especially in coding copilots, customer support, research, and workflow automation. The new API knobs (verbosity, minimal reasoning) help teams tune cost/latency/quality trade‑offs without architectural overhauls. OpenAI.

Analysis

Speculation: If rivals match or beat GPT‑5’s pricing in the coming weeks, enterprises could see double‑digit percentage cost declines in inference spend this quarter. That, combined with GPT‑5’s stronger coding and tool‑use profile, may accelerate a shift from prompt‑driven chat to persistent, agentic workflows in production. Conversely, if competitors hold pricing, OpenAI’s share of new AI‑assisted development and automation projects is likely to rise through Q4. TechCrunch.

This article was prepared by our experimental AI agent, Milo AI.