OpenAI launches GPT-6.1 Sol with near-Astra performance

Investing.com -- OpenAI released GPT-6.1 Sol on Tuesday, a new artificial intelligence model that delivers performance close to its GPT-6 Astra model at one-fifth of the cost. The upgraded model charges $2 per million input tokens and $10 per million output tokens, compared to Astra's $10 and $50 rates respectively.

The company said cached input costs $0.10 per million tokens, representing a 95% reduction from standard input pricing and 50% less than the original GPT-6 Sol's cached input rate.

GPT-6.1 Sol matched GPT-6 Astra's performance on the DeepSWE v1.1 software engineering benchmark while maintaining the lower cost structure. The model scored 6.4 percentage points higher than GPT-6 Sol on the same evaluation at reduced reasoning effort and cost.

On GDP.pdf, which tests models on professional questions using complex documents, GPT-6.1 Sol outperformed Opus 5.5 with fallbacks at less than half the cost per task. The model also scored 2.2 percentage points above Opus 5.5 on AutomationBench at medium reasoning effort, at roughly one-third of the cost.

For computer use tasks, GPT-6.1 Sol scored seven percentage points higher than GPT-6 Sol on OSWorld 2.0's offline set at maximum reasoning effort while costing less than half as much. The model came within 2.1 percentage points of Astra's score at one-seventh the cost per task.

On Terminal-Bench Science 0.1, GPT-6.1 Sol more than doubled GPT-6 Sol's score at maximum reasoning effort. The model averaged $5.47 per task compared to $23.21 for Opus 5.5 and $23.80 for Astra.

The model reduced factual errors from 11.4% to 7.7% at low reasoning effort compared to GPT-6 Sol, a 32% reduction. OpenAI noted these prompts were deliberately difficult and not representative of typical usage.

GPT-6.1 Sol became available Tuesday to Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. Developers can access it through the OpenAI API as gpt-6.1-sol. OpenAI said it will offer GPT-6.1 Sol Ultrafast in the coming days, featuring up to eight times faster token generation.

You May Also Be Interested In

• Gold Fields weighs higher bid for Northern Star Resources - Bloomberg
• Mizuho Reiterates Outperform Rating on Fair Isaac (FICO)
• Eversource, Dartmouth selected for DOE grid technology award