Skip to main content

OpenAI's GPT-6.1 Sol Now Handles 500K Tokens, No Extra Cost

OpenAI's GPT-6.1 Sol Gets a Bigger Brain: 500K Tokens by Default, No Extra Charge

Ever felt like your AI assistant has the memory of a goldfish? You paste in a lengthy report, and halfway through, it forgets what you said at the start. Well, OpenAI just made that problem a lot smaller.

On October 12, OpenAI engineer Thibault Sottiaux (known as Tibo) announced that the default context length for GPT-6.1 Sol has jumped from 272,000 tokens to 500,000 tokens in both Codex and Work environments. That's an 80% increase—practically doubling the amount of text the model can juggle in one go.

What This Means for You

In plain English: your AI can now handle roughly twice as much information before it starts losing track. Whether you're feeding it an entire codebase, a 300-page contract, or a multi-step project with constantly shifting requirements, the conversation is far less likely to hit automatic compression or get chopped off mid-thought.

Think about tracing a clause through a massive legal document, or having AI follow along as your development project grows new features every week. A longer memory means fewer "wait, what were we talking about?" moments and less time re-explaining yourself. For anyone who relies on long-running chats, the improvement is immediate and tangible.

You Can Push It Even Further—Up to 1 Million Tokens

Tibo explained that 500k is the "sweet spot" the team landed on after extensive testing—balancing performance with practical experience for most users. But it's not the ceiling. You can still manually adjust the context window, and GPT-6.1 Sol now supports up to a whopping 1 million tokens.

The Best Part: No Extra Fees for Long Context

Here's where it gets really interesting. Along with the default increase, OpenAI is no longer charging separately for long context. In the past, such expansive windows were treated as premium add-ons, often with a hefty price tag. Now, with the default raised and no additional fees, the barrier for heavy usage just got a lot lower.

It's a smart competitive move, too. Long context has become a must-have selling point across the AI industry, and by passing the benefits directly to users, OpenAI is making a clear play to keep its edge.

Key Points

  • Default context window for GPT-6.1 Sol in Codex and Work environments increased from 272k to 500k tokens—an 80% jump.
  • Manual adjustment still available, with a maximum limit of 1 million tokens.
  • No extra charge for long context, removing a previous cost barrier.
  • Real-world impact: fewer interruptions, less repetition, and better handling of long documents and complex projects.
  • Industry context: Long context is a key battleground, and OpenAI is using it to attract power users.