OpenAI confirms GPT‑6.1 Sol Ultrafast will arrive soon, but pricing and date remain unclear

OpenAI indicated that the faster Ultrafast variant of its GPT‑6.1 Sol model will be released soon.

The confirmation came from Tibo Sottiaux, OpenAI’s product and engineering leader, who replied “6.1 coming soon” to a developer’s query on X on October 4.

The exchange referenced the Ultrafast speed tier that OpenAI had said would arrive “in the coming days” when it launched GPT‑6.1 Sol on September 29.

Model Launch and Pricing

GPT‑6.1 Sol was made available on September 29 for ChatGPT Work, Codex and API users.

OpenAI describes Sol as delivering performance close to its flagship GPT‑6 Astra while costing less for coding, computer‑use and other professional workloads.

The standard API rates for Sol are $2 per million input tokens and $10 per million output tokens, with cached input priced at $0.10 per million tokens.

OpenAI has not published a price for a GPT‑6.1 Sol Ultrafast configuration in the launch announcement.

The launch post promised that the Ultrafast option would provide up to eight times faster token generation than standard speed in Codex.

The “up to eight times” claim represents a maximum, not a guarantee that every prompt or workflow will achieve that multiplier.

The Ultrafast tier is positioned as a separate speed option for users willing to pay more for reduced waiting times.

OpenAI’s API documentation lists Sol’s standard pricing and a 1.05‑million‑token context window.

The Ultrafast guide describes the speed tier for GPT‑6 Astra but does not include a configuration or price for Sol Ultrafast.

Ultrafast Speed Tier Uncertainty

Because the Ultrafast price has not been disclosed, developers cannot compare its cost against the standard Sol rates.

Without a public date for activation, teams are unable to plan deployments that rely on lower latency.

Sottiaux, who joined OpenAI in 2024 after roles at Google and DeepMind, helped build the Codex software‑engineering agent and now fields developer questions.

His brief reply on X offers a timing hint but no concrete guidance on when the mode can be turned on.

OpenAI originally set an expectation of “the coming days” on September 29, then five days later reiterated that the feature was “coming soon” without narrowing the window.

The model itself is already usable, but the promised speed boost remains inaccessible until the Ultrafast tier is priced and released.

This leaves developers in a position where they can test Sol’s capabilities but must wait to assess the latency‑cost trade‑off.

The lack of pricing also prevents cost‑benefit analyses for workloads that might benefit from faster token generation.

For enterprises that prioritize rapid response in coding assistance or interactive applications, the uncertainty may affect budgeting decisions.

Conversely, users satisfied with the standard speed may continue with Sol at its known rates, foregoing the potential latency advantage.

The situation illustrates a broader pattern where model improvements are announced ahead of concrete availability details.

Observers note that OpenAI’s communication style often provides high‑level promises before finalizing pricing structures.

Developers monitoring the rollout should watch OpenAI’s official channels for a formal Ultrafast pricing announcement.

Once the rate is disclosed, they can calculate the incremental spend required for the claimed eight‑fold speed increase.

At that point, they will be able to decide whether the performance gain justifies the additional expense for their specific use cases.

Until then, the Ultrafast tier remains a tentative option rather than an actionable feature.

Why This Matters developers cannot evaluate the latency benefit of GPT‑6.1 Sol Ultrafast against its unknown cost, limiting informed adoption decisions.