OpenAI Releases GPT-5.6 (Sol, Terra, Luna): A Three-Tier Model Family With Programmatic Tool Calling in the Responses API
Evolving story · 6 updatesOpenAI GPT-5.6 LaunchTimeline →OpenAI released GPT-5.6 on July 9, 2026, featuring three pricing tiers and a new programmatic tool calling feature. Sol leads with the highest performance benchmarks.
- GPT-5.6 introduces a three-tier pricing model (Sol, Terra, Luna) with varying performance and cost levels.
- Programmatic Tool Calling enables JavaScript execution in an isolated V8 runtime, reducing token usage and improving efficiency.
- Sol achieves top-tier performance benchmarks, outperforming competitors in coding and multimodal tasks.
- Early adopters report significant token savings, with Clio cutting prompt tokens by 38% and PlayCo by 63.5%.
OpenAI has made GPT-5.6 generally available, marking a significant shift from its previous single-model approach. The new lineup includes three tiers: Sol, Terra, and Luna, each priced differently to cater to varying use cases and budgets. Sol, the premium tier, is priced at $5/$30 per 1M tokens and sets a new benchmark in coding and multimodal tasks. Terra and Luna offer more affordable options at $2.50/$15 and $1/$6 per 1M tokens, respectively.
The standout feature of GPT-5.6 is Programmatic Tool Calling, which allows developers to run model-written JavaScript in an isolated V8 runtime. This enables the model to orchestrate tools without returning every intermediate result, leading to substantial improvements in efficiency. Early adopters like Clio and PlayCo report token savings of 38% and 63.5%, respectively, highlighting the feature's potential to reduce costs and latency.
Performance benchmarks further underscore the model's advancements. Sol achieves an 80 on the Artificial Analysis Coding Agent Index, surpassing Claude Fable 5 by 2.8 points, and scores 62.6% on OSWorld 2.0 while using 85% fewer output tokens than Opus 4.8. These results suggest that GPT-5.6 could redefine expectations for AI-driven coding and multimodal applications.
Programmatic Tool Calling and improved benchmarks make GPT-5.6 a compelling choice for building advanced AI applications.
The tiered pricing model allows businesses to optimize costs while leveraging cutting-edge AI capabilities.
OpenAI's continued innovation in model performance and efficiency signals strong competitive positioning in the AI market.
- Programmatic Tool Calling
- A feature that allows models to execute JavaScript code in an isolated runtime to orchestrate tools without returning intermediate results.
- Artificial Analysis Coding Agent Index
- A benchmark measuring the performance of AI coding agents.
- OSWorld 2.0
- A benchmark evaluating AI models on multimodal and real-world tasks.
LLMOpenAI's GPT-5.6 and ChatGPT Work aim to beat Anthropic on price, speed, and productivity
LLMGPT-5.6 Sol nearly matches Fable 5 on aggregated benchmarks at one-third the cost
LLMOpenAI rolls out GPT-5.6 after government green light — and announces ‘ChatGPT Work’
LLMGPT-5.6: Frontier intelligence that scales with your ambition
LLMNVIDIA Releases Nemotron-Labs-3-Puzzle-75B-A9B: A Compressed Hybrid MoE LLM Delivering 2.03x Server Throughput at Matched User Throughput
BusinessAn AI agent startup just let its agent run its $100 million fundraise
Lyzr, an enterprise AI agent startup, autonomously managed its $100 million funding round using its own AI agent, demonstrating the product's real-world viability.
BusinessOpenAI is shutting down Atlas, but its AI browser ambitions are still growing
OpenAI is retiring its Atlas AI browser after under a year, shifting its agentic browsing features to the desktop app and a Chrome extension.
BusinessElon Musk praises Mythos/Fable, promises not to ‘cut off’ Anthropic
Elon Musk publicly endorses Anthropic and guarantees continued support for hosting its AI models, despite past tensions.
BusinessCan AI answer the $3 trillion question?
A new TechCrunch analysis examines the escalating debate over AI's return on investment, now projected at $3 trillion, and its uneven distribution across industries.
AI Adoption Brings Client, Adviser Returns, per Studies - planadviser
Studies show that AI adoption leads to increased client retention and adviser returns in the financial planning industry.
SecurityPatch for Windows Defender 0-day could allow attackers to fill hard disk
A zero-day vulnerability in Windows Defender could let attackers fill hard drives, crashing systems. Microsoft has released a patch to address the issue.