OpenAI Adds GPT-6.1 Sol Pro, a High-Reasoning Mode of GPT-6.1 Sol, at $2/$10 per 1M Tokens

OpenAI has released GPT-6.1 Sol Pro, a variant of its GPT-6.1 Sol model configured with reasoning.mode set to pro. The listing appeared on OpenRouter on September 29, 2026, priced at $2 per 1M input tokens and $10 per 1M output tokens, with a 1.1 million token context window.

What it is

GPT-6.1 Sol Pro is not a separately trained model. According to OpenAI, it uses the identical underlying weights as GPT-6.1 Sol but runs with a higher reasoning effort setting intended for complex, high-stakes queries. The company states that pro mode consumes substantially more reasoning tokens per request, meaning a typical query costs several times more than the same request sent to standard GPT-6.1 Sol and takes noticeably longer to complete.

OpenAI recommends GPT-6.1 Sol Pro only for hard problems where the added accuracy justifies the cost premium, and continues to recommend standard GPT-6.1 Sol for everyday coding, agentic, and chat workloads.

Pricing and performance data

Across OpenRouter's provider endpoints, pricing and latency vary:

• Azure: $2.00 / $10.00 per 1M tokens (input/output), $0.10 per 1M cached tokens, 2.47s P50 latency, 17 tokens/sec throughput
• OpenAI (direct): $2.00 / $10.00 per 1M tokens, $0.10 per 1M cached tokens, 2.92s P50 latency, 91 tokens/sec throughput, 100% uptime
• OpenAI Flex: $1.00 / $5.00 per 1M tokens (not used in standard routing)
• OpenAI Fast: $4.00 / $20.00 per 1M tokens, 12.49s latency, 99 tokens/sec throughput (not used in standard routing)

Overall availability over the trailing 24 hours was reported at 98.61%, matching the non-routed figure, indicating no failover events during that window.

Context in the GPT-6 lineup

GPT-6.1 Sol Pro sits within OpenAI's broader GPT-6 series, which includes the flagship GPT-6 Astra, the mid-tier GPT-6/6.1 Sol, and the fast, low-cost GPT-6 Luna. Each of these Sol- and Luna-tier models is also offered in a matching "Pro" configuration that applies the same reasoning-mode toggle rather than introducing new weights. Standard GPT-6.1 Sol is priced at $1/$5 per 1M tokens, meaning the Pro configuration roughly doubles per-token cost before accounting for the additional reasoning tokens consumed.

What this means

This is a pricing and configuration change, not a new model release. OpenAI is exposing an existing reasoning-effort parameter as a distinct, separately priced catalog entry rather than shipping new weights. For teams building on GPT-6.1 Sol, the practical decision is a cost-accuracy tradeoff: Pro mode multiplies token spend and latency in exchange for claimed accuracy gains on harder problems, with no new context window, modality, or benchmark data disclosed to quantify that gain. Buyers evaluating GPT-6.1 Sol Pro should benchmark it directly against standard GPT-6.1 Sol on their own workloads before committing to the higher per-token cost.