Model comparison
GPT-5.6 Sol vs GPT-5.5: new cache write fee, lower rates
GPT-5.6 Sol cuts GPT-5.5's rates but adds a 1.25x charge for cache writes. Why the session gap is 16%, not 20%, and what changes on October 14.
· Prices as of September 28, 2026
GPT-5.6 Sol
OpenAI · Released July 9, 2026 · Previous generation
The flagship of the July 2026 GPT-5.6 family, now succeeded by GPT-6 Sol and on promotional pricing.
GPT-5.6 Sol facts and comparisonsGPT-5.5
OpenAI · Released April 23, 2026 · Previous generation
OpenAI's April 2026 flagship for coding and professional work, now a previous generation that stays in the API.
GPT-5.5 facts and comparisons
The short answer
GPT-5.6 Sol charges 20% less than GPT-5.5 for input and 33% less for output, but it bills cache writes at 1.25x input while GPT-5.5 bills them as ordinary input. On the example agentic coding session that narrows the saving to 16%, $4.20 against $5.00. GPT-5.5 leaves ChatGPT and Codex sign-in on October 14, 2026, and GPT-5.6 Sol runs on promotional rates at least through November 21, 2026.
Choose GPT-5.6 Sol if
- You want lower rates on output-heavy work, where the example generation costs $1.72 on GPT-5.6 Sol against $2.55.
- You want explicit cache breakpoints, which OpenAI supports from GPT-5.6 on.
- You value the token efficiency and frontend design gains OpenAI claims for GPT-5.6.
- You use Codex with ChatGPT sign-in, where GPT-5.5 stops being available on October 14, 2026.
Choose GPT-5.5 if
- You call the API directly and have code tuned to GPT-5.5, which stays in the API.
- You want rates with no promotional end date, since GPT-5.6 Sol's are promotional and listed only through November 21, 2026.
- You rely on OpenAI's description of GPT-5.5 for complex coding that needs planning, tool use, and codebase navigation, and have not validated a newer model.
Side by side
Specs and prices
| Fact | GPT-5.6 Sol | GPT-5.5 |
|---|---|---|
| Maker | OpenAI | OpenAI |
| API model id | gpt-5.6-sol | gpt-5.5 |
| Released | July 9, 2026 | April 23, 2026 |
| Status | Previous generation | Previous generation |
| Context window | 1.05M tokens | 1.05M tokens |
| Max output | 128K tokens | 128K tokens |
| Open weights | No | No |
| Input, per 1M tokens | $4 | $5 |
| Cache hit, per 1M | $0.40 | $0.50 |
| Cache write, per 1M | $5 | $5 (same as input) |
| Output, per 1M tokens | $20 | $30 |
| Runs in | Codex, Cursor, OpenCode, OpenRouter, and GitHub Copilot | Codex, Cursor, OpenCode, OpenRouter, and GitHub Copilot |
Standard API rates in US dollars, as published by each maker on September 28, 2026. Batch and priority tiers, taxes, and subscription plans are not included. GPT-5.6 Sol: Requests over 272K input tokens cost 2x for input and cache and 1.5x for output, for the whole request. These are promotional rates, available at least through November 21, 2026. GPT-5.5: Prompts over 272K input tokens cost 2x for input and 1.5x for output, for the full session.
Cost
What the same work costs
The same token counts, priced at each model’s published rates. Cached tokens are billed at each maker’s cache prices, so the session shows what caching is worth on each model.
Real sessions differ: the two models count the same code as different numbers of tokens, and reasoning settings change how much each one writes. Your own history is the real test.
| Workload | GPT-5.6 Sol | GPT-5.5 |
|---|---|---|
| Agentic coding session, 100K input, 400K written to cache, 2M read from cache, 50K output | $4.20 | $5.00 |
| Large one-off review, 150K input with no cache hits, 10K output | $0.80 | $1.05 |
| Output-heavy generation, 30K input, 80K output | $1.72 | $2.55 |
| A month of sessions, 110 sessions: 5 a day, 22 working days | $462.00 | $550.00 |
| Where the session’s cost goes | ||
| Cache writes | $2.00 | $2.00 |
| Cache reads | $0.80 | $1.00 |
| Uncached input | $0.40 | $0.50 |
| Output | $1.00 | $1.50 |
- caching saves on the session with GPT-5.6 Sol (62%)
- $6.80
- caching saves on the session with GPT-5.5 (64%)
- $9.00
Why the gap between GPT-5.6 Sol and GPT-5.5 shrinks on cached sessions
On list prices, GPT-5.6 Sol is cheaper wherever it counts. It charges $4 per million input tokens against $5 for GPT-5.5, $0.40 per million cached tokens against $0.50, and $20 per million output tokens against $30.
The difference lies in how cache writes are billed. From GPT-5.6 on, OpenAI charges 1.25x the input price to write the cache, which makes a GPT-5.6 Sol write cost $5 per million tokens. GPT-5.5 adds no write charge, so its written tokens cost ordinary input, which is also $5. On every written token, the newer model's lower input price is cancelled out.
In the example session 400K tokens go into the cache, and they cost $2.00 on both models. That is 48% of GPT-5.6 Sol's session and 40% of GPT-5.5's. The saving comes only from cache reads, fresh input, and output, which is why the session gap is 16% while the output-heavy generation shows 33%.
How OpenAI's caching rules differ between GPT-5.5 and GPT-5.6
Beyond the write price, OpenAI documents other caching rules for GPT-5.6 and later. You can mark up to four explicit cache breakpoints, while GPT-5.5 and earlier cache automatically only. On GPT-5.6 and later, a cached prefix stays reusable for at least 30 minutes after its last use, and caching starts at 1,024 input tokens. A cache hit costs 0.1x input on both generations.
As a share, caching saves more on GPT-5.5: 64% of the uncached session, against 62% on GPT-5.6 Sol, because GPT-5.5 pays nothing extra to write. In dollars it saves $9.00 on GPT-5.5 and $6.80 on GPT-5.6 Sol. Neither figure makes GPT-5.5 the cheaper model, since its session still costs $0.80 more.
Long prompts follow a similar rule with different wording. Both models raise prices past 272K input tokens, to 2x for input and 1.5x for output. OpenAI scopes that to the whole request on GPT-5.6 Sol and to the full session on GPT-5.5. The example session keeps every request below 272K, so neither rule applies to it.
GPT-5.5 leaves Codex sign-in, and GPT-5.6 Sol already has a successor
GPT-5.5 leaves ChatGPT and Codex sign-in on October 14, 2026, and stays in the API. GPT-5.6 Sol is a previous-generation model too: Codex suggests moving from it to GPT-6 Sol, though it still serves Codex cloud chats on ChatGPT plans.
GPT-5.6 Sol's current rates are promotional, available at least through November 21, 2026, and this page has no source for its price after that. The 16% session gap is today's figure. If you are migrating off GPT-5.5 anyway, price GPT-6 Sol on your own sessions before settling on GPT-5.6 Sol.
OpenAI pitched GPT-5.5 for complex coding that needs planning, tool use, codebase navigation, and verification, with fewer reasoning tokens than earlier models at the same effort. It pitched GPT-5.6 Sol for token efficiency, frontend layout and design judgment, and programmatic tool calling. Their context windows and output limits are identical, 1.05M and 128K. EveryToken splits your local Codex history by model and shows what caching saved or cost on each.
Prompt caching
How OpenAI bills cached tokens
OpenAI
Prompt caching is on by default. From GPT-5.6 on you can also mark up to four explicit cache breakpoints, while GPT-5.5 and earlier cache automatically only.
From GPT-5.6 on, a cache write costs 1.25x the uncached input price and a cache hit costs 0.1x. GPT-5.5 and earlier add no charge for writing the cache: written tokens are billed as ordinary input, and a hit costs 0.1x on the models compared here.
On GPT-5.6 and later, a cached prefix stays reusable for at least 30 minutes after its last use, and caching starts at 1,024 input tokens.
Source: OpenAI: Prompt caching
Your own numbers
See what GPT-5.6 Sol and GPT-5.5 really cost you.
everyaitoken reads your Codex, Cursor, OpenCode, and OpenRouter history on your Mac and prices every request at API rates, with what caching saved or cost. $9 once.
FAQ
Questions
Is GPT-5.6 Sol cheaper than GPT-5.5?
Yes, on every rate except cache writes, which cost $5 per million tokens on both. One example session runs $4.20 on GPT-5.6 Sol and $5.00 on GPT-5.5, and 110 of them a month come to $462.00 against $550.00.
When does GPT-5.5 leave Codex?
On October 14, 2026, for ChatGPT and Codex sign-in. API access continues after that date.
Does GPT-5.5 charge extra for cache writes?
No. On GPT-5.5, written tokens cost the same as ordinary input. OpenAI's 1.25x write charge starts with GPT-5.6.
How long do GPT-5.6 Sol's promotional rates last?
OpenAI lists them as available at least through November 21, 2026. The price that follows is not in the sources this page uses.