OpenAIOfficial ReleaseOfficial Source

Advancing the price-performance frontier with GPT-5.6

OpenAI says GPT-5.6 Luna and Terra now cost less, which may change model-routing choices for agent workflows. The supplied material gives no prices or workload comparisons.

OpenAI · Jul 30, 2026

Open Source Open Markdown Open JSON

Source Summary

OpenAI reports **lower pricing** for **GPT-5.6 Luna** and **GPT-5.6 Terra**, attributing the change to improved model efficiency.

Practical Implication

Revisit model-routing and cost assumptions for sustained agent workloads, but calculate the effect using your own request mix and tool-call patterns.

Agent-Ready Context

OpenAI reports **lower pricing** for **GPT-5.6 Luna** and **GPT-5.6 Terra**, attributing the change to improved model efficiency.

Revisit model-routing and cost assumptions for sustained agent workloads, but calculate the effect using your own request mix and tool-call patterns.

The supplied material includes no prices, benchmark results, or workload-level comparisons, so the practical savings cannot be quantified here.

Connected Context · Feed7 Judgment

This turns GPT-5.6’s broad efficiency positioning into a concrete routing signal for Luna and Terra: their lower prices warrant recalculating sustained agent costs. It still does not show whether either tier beats alternatives for a particular workload, so request mix, tool use, and observed quality remain necessary inputs.

How GPT-5.6 fuses frontier intelligence with frontier efficiencyThe price reduction gives a concrete basis for part of the earlier useful-output-per-dollar claim, while still leaving workload-level gains unmeasured.GPT 5.6 Sol, Luna, and Terra now available on AI GatewayThe gateway exposes Luna and Terra as routing targets, so their revised prices can directly affect tier selection within that deployment path.Introducing Cursor RouterCursor’s task-and-cost routing approach is an implementation consequence of changing model economics: lower Luna and Terra prices may alter routing thresholds, but must be tested on local traffic.CFOs and the new economics of AIThe reported variation in cost per agent request reinforces why lower list pricing should be translated through actual model mix and usage rather than treated as a uniform saving.

Context Map

model#model-selection#enterprise

Uncertainty

The supplied material includes no prices, benchmark results, or workload-level comparisons, so the practical savings cannot be quantified here.