OpenAIOpenAI·💬 Text Generation·↑ Newer: GPT-5.6 Sol·New

GPT-6 Sol

ReasoningVisionFunction CallingWeb Searchanonymized
🧠 Try in Intelligence →Try on Venice.ai ↗
Quick reference
GPT-6 Sol — TLDR
  • 🧠 OpenAI's GPT-6 coding and agentic workflow model, released September 2026
  • 📏 1.05M-token context window with up to 128K output tokens
  • 🔧 Reasoning effort configurable from none through low, medium, high, xhigh, max
  • 👁️ Accepts text and image inputs, generates text
  • 🎯 Function calling, structured outputs and Responses API built-in tools
  • ⚡ OpenAI reports substantial gains over GPT-5.6 Sol on FrontierCode
  • 🏢 Positioned as the intelligence-and-cost balance tier between Astra and Luna
💰 Pricing
$2.50 / $12.50
per 1M · input / output
📏 Context
1.1M tokens
📅 On Venice since
Sep 22, 2026
1 day ago
Provider

OpenAI is an American artificial intelligence research organization headquartered in San Francisco, structured as both a for-profit public benefit corporation and a nonprofit foundation. The lab developed the GPT family of large language models, the DALL-E…

Read full profile →
38 models on Venice
23 text · 4 image · 4 video · 4 inpaint · 2 embedding · 1 asr
Since Jan 15, 2025

About this model

GPT-6 Sol is OpenAI's GPT-6-generation model built for complex coding and agentic workflows, launched alongside GPT-6 Luna on 22 September 2026. In OpenAI's own guidance, GPT-6 Astra is the flagship for the hardest reasoning and coding, Sol is the choice for balancing intelligence and cost, and Luna targets cost-sensitive, high-volume workloads.

Technically, Sol handles a 1,050,000-token context window with up to 128,000 output tokens, accepts text and image inputs, and exposes a reasoning-effort dial spanning none, low, medium (default), high, xhigh and max. Structured outputs, function calling and the Responses API built-in tools make it suited to multi-step automation and long-horizon repository work.

Compared with its direct predecessor, GPT-5.6 Sol — which carried the same 1.05M context window and the same reasoning-effort settings — OpenAI reports that GPT-6 Sol improves substantially on FrontierCode, an evaluation of whether coding agents produce changes ready to merge into real codebases. OpenAI also publishes self-reported figures of 56.4% on Agents' Last Exam at max effort, 68.8% on DeepSWE v1.1 at max effort, and 60.5% on OSWorld 2.0 offline at xhigh effort.

OpenAI states that the release focuses on capability improvements and cost efficiency for Sol and Luna, accompanied by higher usage limits.

This About section is AI-generated from public sources (Claude Opus 5), with no human editing. It may contain inaccuracies — verify critical details against the sources listed above.

Data sources: Venice API · HuggingFace · Wikipedia — enrichment updated 2h ago