ATRIUMsearch → argument graph
Article · 2026-07-30 · 4 moments

Advancing the price-performance frontier with GPT‑5.6

Advancing the price-performance frontier with GPT‑5.6 Huge price drop from OpenAI today: GPT-5.6 Terra got a 20% reduction, and GPT-5.6 Luna got a massive 80% drop. OpenAI credit 5.6 Sol with enabling this: in How GPT‑5.6 fuses frontier intelligence with frontier efficiency they describe using 5.6 Sol to optimize load balancing, and more impressively to optimize inference itself: We also used GPT‑5.6 Sol to optimize the model’s forward pass: the computation that transforms inputs into next-to ✦ AI generated

Read original ↗
GUEST OpenAIGUEST AuthorHOST Simon Willison
01
Data

At $0.20/$1.20 per million tokens, GPT-5.6 Luna is cheaper than Google's Gemini 3.1 Flash-Lite and 1/5th the input price of Anthropic's Claude Haiku 4.5.

The author compares Luna's new pricing to Google's Gemini 3.1 Flash-Lite and Anthropic's Claude Haiku 4.5, finding Luna significantly cheaper than both.

transcript

Author: That Luna price drop completely changes the landscape with respect to lower priced models. At $0.20/million tokens for input and $1.20/million for output Luna is now cheaper than Google's Gemini 3.1 Flash-Lite ($.025/$1.50). Anthropic's cheapest current model is Claude Haiku 4.5, and that's $1/$5 - Luna is now 1/5th of that for input, previously it cost the same.

explains mechanism · 1provides context · 1

03
Mechanism◆

OpenAI used GPT-5.6 Sol to optimize the model's forward pass and autonomously rewrite production kernels in Triton and Gluon, reducing end-to-end serving costs by 20%.

OpenAI explains how GPT-5.6 Sol optimized load balancing and the computational forward pass, using Codex to rewrite GPU kernels and reduce serving costs by 20%.

transcript

OpenAI: We also used GPT‑5.6 Sol to optimize the model's forward pass: the computation that transforms inputs into next-token predictions. Even when individual operations are fast, excess memory movement, synchronization, and inefficient data layouts can leave GPUs idle. To avoid this, GPT‑5.6 Sol found work that could be precomputed, avoided, or parallelized. With Codex, GPT‑5.6 Sol autonomously rewrote and optimized our production kernels, the core code that executes the mathematical operations that make up the model. This worked in part because we've trained GPT‑5.6 to be effective at writing and improving kernels in Triton and Gluon, two open-source GPU programming languages maintained by OpenAI. These efforts, combined with broader kernel advancements from GPT‑5.6 Sol, reduced end-to-end serving costs by 20%.

Highlight slides
OpenAI announces GPT-5.6 price cuts✦ from: OpenAI reduced GPT-5.6 Terra pricing by 20% and GPT-5.6 Luna pricing by 80%.Forward Pass Optimization by GPT-5.6 Sol✦ from: OpenAI used GPT-5.6 Sol to optimize the model's forward pass and autonomously rewrite production kernels in Triton and Gluon, reducing end-to-end serving costs by 20%.Luna cut dwarfs Terra reduction✦ from: OpenAI reduced GPT-5.6 Terra pricing by 20% and GPT-5.6 Luna pricing by 80%.20% Serving Cost Reduction via Kernel Rewriting✦ from: OpenAI used GPT-5.6 Sol to optimize the model's forward pass and autonomously rewrite production kernels in Triton and Gluon, reducing end-to-end serving costs by 20%.
Related episodes
Article · Latent Space

[AINews] OpenAI launches GPT 5.6 Sol/Terra/Luna, Codex becomes ChatGPT superapp

→
Article · Big Technology

Apple’s lawsuit against OpenAI makes serious claims. Will they matter?

→
Video · The Cognitive Revolution

Fable's Back, AI Engineer Recap, & SambaNova

→