name = "GPT-6 Sol"
description = "OpenAI model for complex coding and agentic workflows"
family = "gpt-sol"
release_date = "2026-09-22"
last_updated = "2026-09-22"
attachment = true
reasoning = true
temperature = false
tool_call = true
structured_output = true
knowledge = "2026-04-20"
open_weights = false

[limit]
context = 1_050_000
input = 922_000
output = 128_000

[modalities]
input = ["text", "image", "pdf"]
output = ["text"]

[[benchmarks]]
name = "DeepSWE"
score = 68.8
metric = "score"
version = "1.1"
variant = "max effort"
source = "https://openai.com/index/introducing-gpt-6-sol-and-luna/"
date = "2026-09-22"

[[benchmarks]]
name = "AutomationBench"
score = 33.2
metric = "score"
version = "1.0.6"
variant = "xhigh effort"
source = "https://openai.com/index/introducing-gpt-6-sol-and-luna/"
date = "2026-09-22"

[[benchmarks]]
name = "Agents' Last Exam"
score = 56.4
metric = "score"
version = "1"
variant = "max effort"
source = "https://openai.com/index/introducing-gpt-6-sol-and-luna/"
date = "2026-09-22"

[[benchmarks]]
name = "OSWorld"
score = 60.5
metric = "partial reward"
version = "2.0"
dataset = "offline v2026.08.08"
variant = "xhigh effort"
source = "https://openai.com/index/introducing-gpt-6-sol-and-luna/"
date = "2026-09-22"