GPT-5.6 Wins Early Praise on Price and Power

GPT-5.6 has received an unusually favorable initial reception. OpenAI's new Sol, Terra, and Luna model family is drawing praise not just for benchmark gains, but for pairing frontier performance with pricing that makes deployment practical. Performance and cost efficiency are therefore at the center of early evaluations.

The impact is most pronounced in coding. Independent evaluations place Sol near the top of the general intelligence field and at the head of a key coding-agent index. OpenAI is pairing those gains with repeated usage resets to accelerate user growth. While OpenAI has not yet proved it has overtaken Anthropic, the competition has become significantly tighter.

GPT-5.6 Pricing Turns Benchmarks Into a Product Story

This shifting dynamic is driven by the metrics. OpenAI's launch page introduces GPT-5.6 as a three-model family, while Artificial Analysis scored Sol at 59 on its Intelligence Index, just one point behind Claude Fable 5. In those tests, Sol cost $1.04 per task—approximately one-third of Fable 5's cost—and scored 80 on the Coding Agent Index when paired with Codex.

The model family offers developers a clear pricing tier. Artificial Analysis lists Sol, Terra, and Luna at $5/$30, $2.50/$15, and $1/$6 per million input/output tokens, respectively. At maximum reasoning, Terra and Luna cost roughly 50% and 80% less per Intelligence Index task than Sol. While these figures do not reflect all real-world workloads, they provide a basis for evaluating GPT-5.6's balance of performance and affordability.

Seven Million Users and a Drumbeat of Codex Resets

On July 12, OpenAI product lead Tibo Sottiaux announced on X that Codex and ChatGPT Work had reached 6 million active users after an intense 48 hours. OpenAI temporarily removed the five-hour usage-limit restriction for Plus, Business, and Pro plans and said another usage reset would arrive within the hour.

Sottiaux subsequently announced that Codex and ChatGPT Work had surpassed 7 million active users. OpenAI commemorated the milestone by introducing a "banked reset" in its desktop app and web interface, allowing users to manually replenish weekly usage.

Anthropic Extends Two Access Promotions

Anthropic has extended two separate promotions through July 19, 2026. Its official Claude Code offer raises weekly limits by 50% for Pro, Max, Team, and legacy seat-based Enterprise users. The increase applies automatically across the command-line interface, IDE extensions, desktop app, and web, while the standard five-hour limit remains unchanged.

In addition, Anthropic extended its promotional Fable 5 access for paid subscribers through July 19. Anthropic said eligible users can use up to 50% of their weekly subscription limits on Fable 5 at no additional cost. Industry observers may interpret the extensions as Anthropic weighing its next response to intensifying competition, but the company has not said they were timed as a direct reaction to OpenAI. As earlier AIScroll coverage noted, model access, package value, and operating costs are now central to comparisons between the platforms.

A Shifting Coding Landscape as GPT-6 Rumors Circulate

This competitive pressure has fueled speculation of a market shift. Some developers and industry observers suggest that Codex has nearly caught up to Claude Code, pointing to a potential "golden cross." While this narrative captures current market momentum, it does not reflect verified market share, as no audited comparative data proves OpenAI has overtaken Anthropic in the coding agent space.

Meanwhile, speculative focus is already shifting to the next generation of models. Andrew Curran wrote on X that he had independently heard GPT-6 would be OpenAI's next release, potentially arriving within four weeks. Because OpenAI has not confirmed this timeline or any upcoming launch, the report remains a social media rumor rather than an official product roadmap.

OpenAI was widely seen as trailing after Claude Code launched, but GPT-5.6 and Codex have put it back into a tightening race for the lead. With the gap narrowing quickly and GPT-6 rumors now in the mix, the future balance of power is set to become even more compelling to watch.