OpenAI API pricing

OpenAI's current flagship: deepest reasoning in the 5.6 family with a 1.05M-token context window and Sol-tier throughput.

The balanced 5.6 tier: near-Sol quality at $2/$12 with the full 1.05M context window — the default choice for production workloads.

The cheap 5.6 tier at $0.20/$1.20 — classification, routing and high-volume features with frontier-family latency.

Previous flagship generation, still fully supported at $5/$30 with the 1.05M context window — now effectively superseded by 5.6 Sol pricing.

Mid-generation 5.4 flagship tier at $2.50/$15 with the 1.05M context window.

The 5.4 small tier at $0.75/$4.50 — the workhorse for chat and RAG before the 5.6 Luna price reset.

Cheapest 5.4 tier at $0.20/$1.25 for extraction, tagging and routing at scale.

Popular 2025 generation at $1.75/$14 with 400K context — still common in production stacks.

The $1.25/$10 generation that anchored 2025 budgets, still listed and fully supported.

Dedicated reasoning model that spends thinking tokens before answering — strong on hard math and science.

Compact reasoning model at $1.10/$4.40 — cheap thinking for structured tasks and tool loops.

The long-context veteran with a 1.05M-token window, still listed for existing production pipelines.