Breaking

OpenAI Adds Ultrafast Mode for GPT-6.1 Sol in the Responses API

October 8, 2026 at 15:06 EDT

Foundation Models
AI Agents
Software Dev & Coding

OpenAI has added Ultrafast mode for GPT-6.1 Sol in the Responses API, opening a lower-latency path for its near-Astra model with service_tier: "ultrafast". The move turns the speed claim behind the launch - up to 8x faster token generation than Standard in Codex, or roughly 700% faster - into a paid tier aimed at coding, agents and live applications.

Oct 8, 2026 · OpenAI

GPT-6.1 Sol gets Ultrafast

OpenAI added its lowest-latency service tier for GPT-6.1 Sol, trading higher API cost for faster token generation in coding, agent and live workflows.

up to faster token generation than Standard

$12

per million input tokens in Ultrafast

$60

per million output tokens in Ultrafast

1.05M

token context window

Speed scale

Standard

Ultrafast

up to 8x

The headline gain is speed, not a new model checkpoint.

Where it helps

Interactive coding loops, outage debugging, agents navigating apps and live interfaces where seconds matter.

Where it pinches

API usage is priced at 6x Standard, with separate rate limits teams must manage before scaling traffic.

How the request changes

model: gpt-6.1-sol→service_tier: ultrafast→Responses API

WebSockets are recommended for agentic apps with many quick tool calls.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.