GPT-6.1 Sol — Specs, Pricing & Limits
Specs
| Status | Priced |
|---|---|
| Released (API changelog) | 2026-09-29 |
| Codex CLI | The CLI's current default model |
| Context window | 1,050,000 tokens |
| Max output | 128,000 tokens |
| Knowledge cutoff | Apr 30, 2026 |
| Reasoning effort levels | low, medium (default), high, xhigh, max |
| Long-context rule | >272,000 input tokens: 2x input and cache rates, 1.5x output (full request) |
| Cache writes | Billed at the cache-write rate instead of the input rate |
| ChatGPT credit rate (per 1M) | 50 in · 2.5 cached · 250 out |
Pricing per 1M tokens, every published service tier
| Standard — short context | in $2.00 · cached $0.10 · writes $2.50 · out $10.00 |
|---|---|
| Standard — long context | in $4.00 · cached $0.20 · writes $5.00 · out $15.00 |
| Batch — short context | in $1.00 · cached $0.05 · writes $1.25 · out $5.00 |
| Batch — long context | in $2.00 · cached $0.10 · writes $2.50 · out $7.50 |
| Flex — short context | in $1.00 · cached $0.05 · writes $1.25 · out $5.00 |
| Flex — long context | in $2.00 · cached $0.10 · writes $2.50 · out $7.50 |
| Fast — short context | in $4.00 · cached $0.20 · writes $5.00 · out $20.00 |
| Fast — long context | in $8.00 · cached $0.40 · writes $10.00 · out $30.00 |
Rate limits by usage tier
| Free | Not supported |
|---|---|
| Tier 1 | 500 RPM · 500,000 TPM · batch queue not published |
| Tier 2 | 5,000 RPM · 1,000,000 TPM · batch queue not published |
| Tier 3 | 5,000 RPM · 2,000,000 TPM · batch queue not published |
| Tier 4 | 10,000 RPM · 4,000,000 TPM · batch queue not published |
| Tier 5 | 15,000 RPM · 40,000,000 TPM · batch queue not published |
GPT-6.1 Sol is the model the Codex CLI runs when you don't choose one, and OpenAI's recommendation for complex coding work in Codex where your account has it. OpenAI pitches it as near-Astra performance at a lower cost, and the price table backs the second half: its Standard rates are a fraction of GPT-6 Astra's while sharing Astra's context window, output ceiling and knowledge cutoff.
Its distinguishing price cell is cached input. Every other model with a published cached-input rate bills a cache read at 0.1x its input rate; GPT-6.1 Sol bills it at 0.05x. Agentic Codex sessions re-send large cached contexts on every turn, so this one cell often matters more to a real bill than the headline input price — and it's the main thing separating it from GPT-6 Sol, which matches it on input and output.
A couple of constraints catch migrations. It accepts neither none nor minimal reasoning effort — its range runs from low to max, with medium as the default — and tool calling requires the Responses API. Its published rate-limit table matches GPT-6 Sol's requests and tokens per minute but has no batch-queue column, so no batch queue limit is shown here rather than one borrowed from a sibling.
In Codex, its launch rollout covered Plus, Pro, Business, Enterprise and Edu, with Enterprise and Edu keeping it off until an administrator enables it; Free and Go weren't included. Standard and Fast are available, and OpenAI describes Ultrafast for this model as coming later. GPT-6.1 Sol vs GPT-5.6 Sol compares it with the default it replaced.
The tables above are computed from this site's facts module: every published service tier with both context bands, the effort levels, and the rate-limit table.
Verified 2026-10-01 against https://developers.openai.com/api/docs/models/gpt-6.1-sol.
Could not confirm: Its rate-limit table has no batch-queue column, so no batch queue limit is shown for it — not borrowed from GPT-6 Sol's table. Ultrafast is described as coming later for this model; no Ultrafast price is published for it yet.
Checked: https://developers.openai.com/api/docs/models/gpt-6.1-sol · https://developers.openai.com/api/docs/pricing · https://developers.openai.com/api/docs/changelog · https://learn.chatgpt.com/docs/codex/cli