GPT-6.1 Sol Ultrafast

Kingy AI confirms OpenAI added the Ultrafast service tier to GPT-6.1 Sol in the Responses API on October 8, 2026, with global processing and US/EU data residency for all API users subject to rate limits. The underlying model identifier remains gpt-6.1-sol, and developers select the tier with service_tier: ultrafast. Short-context Ultrafast on GPT-6.1 Sol is priced at $12 input and $60 output per million tokens, a 6× markup over Standard's $2 and $10. In matched Kingy trials, coding median completion dropped from 9.61 to 2.61 seconds and a research task's first response fell from 19.88 to 5.06 seconds, with both tiers delivering equal-quality accepted outputs.

VentureBeat reports OpenAI's new Ultrafast inference tier delivers up to 8× faster token generation in Codex and 6× faster in the API, reaching as much as 300 tokens per second on GPT-6-class frontier models. The speed carries a 6× premium over the underlying model's standard API rate, targeting workloads where latency, not token cost, is the bottleneck. Ultrafast launched immediately for GPT-6 Astra with a 6.1 Sol version coming soon, broadening enterprise price-performance options. Artificial Analysis comparisons show Gemini 3.5 Flash around 201 t/s while speed-optimized models Mercury 2 and Celeris-1 reach roughly 769 and 1,491 t/s respectively, framing OpenAI's 300 t/s as a frontier-tier ceiling.

OpenAI introduced GPT-6.1 Sol on September 29, 2026, as an upgrade to GPT-6 Sol that nearly matches GPT-6 Astra on agentic coding, computer use, and professional work at one-fifth of Astra's standard token prices. Standard pricing is $2 input, $10 output, and $0.10 cached input per million tokens, a 50% cut in cached input from GPT-6 Sol. On DeepSWE v1.1 GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth the cost and beats GPT-6 Sol by 6.4 percentage points. It also edges Claude Opus 5.5 by 2.2 points on AutomationBench and outscores it on GDP.pdf at less than half the per-task cost, positioning Sol as a default for everyday agentic work.

Venice documents the exact served variant openai-gpt-61-sol-ultrafast as OpenAI's GPT-6.1 Sol running in its fastest scheduling tier, live on the platform from October 8, 2026. It delivers near-flagship performance for coding, computer use, and professional work with a 1,050K-token context window, vision input, tool calling, reasoning, and web search all intact. The page lists reasoning efforts slow through max (none and minimal unsupported), prompt length up to 922,000 input tokens, and an April 30, 2026 knowledge cutoff. Venice exposes the model via an OpenAI-compatible API by changing the base URL and model id, billed per token under an anonymized privacy tier that stores no prompts or profiles.