LATEST MODEL
View in the Herd

GPT-6.1 Sol

OpenAI ⚔️ The Knight Released September 2026

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 token prices

GPT-6.1 Sol

OpenAI • September 2026

Latest

Training Data

Not disclosed

GPT-6.1 Sol

September 2026

Parameters

Not disclosed

Training Method

Reinforcement Learning with Chain-of-Thought Reasoning

Context Window

1,050,000 tokens

Knowledge Cutoff

April 30, 2026

Key Features

1M Token Context • Near-Astra at 1/5 the Price • $0.10 Cached Input • Reasoning Effort: Low to Max

Capabilities

Coding: Maximum (matches GPT-6 Astra on DeepSWE v1.1)

Professional Work: Outstanding (beats Opus 5.5 on GDP.pdf and AutomationBench)

Computer Use: Outstanding (+7 pts over GPT-6 Sol on OSWorld 2.0 offline)

What's New in This Version

Matches GPT-6 Astra on DeepSWE v1.1 at ~1/5 the cost and beats GPT-6 Sol's best by 6.4 points; tops Opus 5.5 on AutomationBench by 2.2 points at ~1/3 the cost; comes within 2.1 points of Astra on OSWorld 2.0 offline at ~1/7 the cost per task; cuts factual errors ~32% vs GPT-6 Sol at low effort; $2/$10 per 1M tokens, cached input $0.10 (half of GPT-6 Sol's)

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 token prices

What's New in This Version

Matches GPT-6 Astra on DeepSWE v1.1 at ~1/5 the cost and beats GPT-6 Sol's best by 6.4 points; tops Opus 5.5 on AutomationBench by 2.2 points at ~1/3 the cost; comes within 2.1 points of Astra on OSWorld 2.0 offline at ~1/7 the cost per task; cuts factual errors ~32% vs GPT-6 Sol at low effort; $2/$10 per 1M tokens, cached input $0.10 (half of GPT-6 Sol's)

Technical Specifications

Parameters Not disclosed
Context Window 1,050,000 tokens
Training Method Reinforcement Learning with Chain-of-Thought Reasoning
Knowledge Cutoff April 30, 2026
Training Data Not disclosed

Key Features

1M Token Context Near-Astra at 1/5 the Price $0.10 Cached Input Reasoning Effort: Low to Max

Capabilities

Coding: Maximum (matches GPT-6 Astra on DeepSWE v1.1)
Professional Work: Outstanding (beats Opus 5.5 on GDP.pdf and AutomationBench)
Computer Use: Outstanding (+7 pts over GPT-6 Sol on OSWorld 2.0 offline)
Theme
Language
Support
© funclosure 2025