Model comparison
GPT-5.6 Sol vs Gemini 3.1 Pro Preview: two models in flux
GPT-5.6 Sol is on promotional rates and Gemini 3.1 Pro Preview is still in preview. On a cached coding session, Gemini costs $2.00 against $4.20.
· 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 comparisonsGemini 3.1 Pro Preview
Google · Released February 19, 2026 · Preview
Google's only current Pro model, still in preview, positioned for deep reasoning and agentic coding.
Gemini 3.1 Pro Preview facts and comparisons
The short answer
Gemini 3.1 Pro Preview is cheaper: the example agentic coding session costs $2.00 on it and $4.20 on GPT-5.6 Sol, a 2.1x gap, and 110 sessions a month come to $220.00 against $462.00. Both are in transition, since OpenAI has succeeded GPT-5.6 Sol with GPT-6 Sol and keeps it on promotional rates, while Google has announced Gemini 3.5 Pro. Pick GPT-5.6 Sol for 128K outputs and OpenAI's frontend and tool-calling strengths, and Gemini 3.1 Pro Preview for lower cost in Gemini CLI.
Choose GPT-5.6 Sol if
- You need up to 128K output tokens, against 65.5K on Gemini 3.1 Pro Preview.
- Frontend work matters to you, and OpenAI claims better layout, visual hierarchy, and design judgment for GPT-5.6.
- Your Codex cloud chats on a ChatGPT plan already run on GPT-5.6 Sol.
- You want programmatic tool calling, which OpenAI describes as the model writing JavaScript that calls tools and processes their output.
Choose Gemini 3.1 Pro Preview if
- Price comes first: $2 input and $12 output per million tokens, against $4 and $20.
- You work in Gemini CLI, which uses Gemini 3.1 Pro Preview for the Pro half of its default auto model.
- Your sessions write the cache often, since Google bills written tokens as ordinary input while OpenAI charges 1.25x for GPT-5.6 Sol.
Side by side
Specs and prices
| Fact | GPT-5.6 Sol | Gemini 3.1 Pro Preview |
|---|---|---|
| Maker | OpenAI | |
| API model id | gpt-5.6-sol | gemini-3.1-pro-preview |
| Released | July 9, 2026 | February 19, 2026 |
| Status | Previous generation | Preview |
| Context window | 1.05M tokens | 1.05M tokens |
| Max output | 128K tokens | 65.5K tokens |
| Open weights | No | No |
| Input, per 1M tokens | $4 | $2 |
| Cache hit, per 1M | $0.40 | $0.20 |
| Cache write, per 1M | $5 | $2 (same as input) |
| Output, per 1M tokens | $20 | $12 |
| Runs in | Codex, Cursor, OpenCode, OpenRouter, and GitHub Copilot | Cursor, Gemini CLI, OpenCode, and OpenRouter |
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. Gemini 3.1 Pro Preview: Prompts over 200K input tokens cost $4 input, $0.40 cached, and $18 output per million tokens.
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 | Gemini 3.1 Pro Preview |
|---|---|---|
| Agentic coding session, 100K input, 400K written to cache, 2M read from cache, 50K output | $4.20 | $2.00 |
| Large one-off review, 150K input with no cache hits, 10K output | $0.80 | $0.42 |
| Output-heavy generation, 30K input, 80K output | $1.72 | $1.02 |
| A month of sessions, 110 sessions: 5 a day, 22 working days | $462.00 | $220.00 |
| Where the session’s cost goes | ||
| Cache writes | $2.00 | $0.80 |
| Cache reads | $0.80 | $0.40 |
| Uncached input | $0.40 | $0.20 |
| Output | $1.00 | $0.60 |
- caching saves on the session with GPT-5.6 Sol (62%)
- $6.80
- caching saves on the session with Gemini 3.1 Pro Preview (64%)
- $3.60
Where the $2.20 session gap comes from
GPT-5.6 Sol charges 2x the input and cache hit prices of Gemini 3.1 Pro Preview, $4 against $2 and $0.40 against $0.20, and 1.7x its output price, $20 against $12. Cache writes are further apart, 2.5x, because OpenAI charges 1.25x input for a write from GPT-5.6 on, $5 per million, while Google bills written tokens as $2 input.
In the example session, writes account for $1.20 of the $2.20 difference: $2.00 on GPT-5.6 Sol against $0.80 on Gemini. Reads add $0.40, input $0.20, and output $0.40. The uncached review shows a smaller 1.9x gap, $0.80 against $0.42, and output-heavy generation 1.7x, $1.72 against $1.02.
A promotional price and a preview label
Neither model's current terms look settled for the long run. OpenAI lists GPT-5.6 Sol's rates as promotional, available at least through November 21, 2026, and has succeeded it with GPT-6 Sol. Codex cloud chats on ChatGPT plans still run on GPT-5.6 Sol, and the API id gpt-5.6 points to it, but elsewhere Codex suggests GPT-6 Sol.
On Google's side, Gemini 3.1 Pro Preview still carries the preview label it launched with on February 19, 2026. Gemini 3.5 Pro is announced but unreleased, and GitHub Copilot dropped the preview model on September 1, 2026. A team planning a year ahead should expect to revisit either choice.
Output limits, long prompts, and tool calling
GPT-5.6 Sol writes up to 128K tokens per response and Gemini 3.1 Pro Preview up to 65.5K. Both list a 1.05M context window. Google charges $4 input, $0.40 cached, and $18 output once a prompt passes 200K input tokens, while OpenAI's long-context rule starts at 272K and bills the whole request at 2x for input and cache and 1.5x for output.
Each maker has its own approach to agent tools. OpenAI credits GPT-5.6 with programmatic tool calling, where the model writes JavaScript that calls tools and processes their output. Google offers a separate customtools endpoint for Gemini 3.1 Pro Preview that it says prioritizes custom tools alongside bash, and Gemini CLI gives it to Gemini API key users at the same price.
Gemini's thinking cannot be turned off and defaults to high, and OpenAI says GPT-5.6 reaches flagship-level performance with fewer output tokens. Both change real output counts, which the fixed-token cost table cannot reflect. EveryToken prices your actual Codex and Gemini CLI sessions at API rates, per model.
Prompt caching
How each maker 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
Implicit caching is on by default for Gemini 2.5 and newer. When a request repeats a prefix Google has cached, the discount is applied automatically, but a hit is not guaranteed.
Explicit caching creates a cache you reference by name, with a guaranteed discount. It adds a storage charge for as long as the cache lives, 1 hour by default: $4.50 per million tokens per hour on Pro models and $0.50 to $1 on Flash models.
On every model compared here, a cache hit costs 10% of the input price. Google publishes no separate cache-write price, so this blog prices written tokens as ordinary input.
Source: Google: Context caching
Your own numbers
See what GPT-5.6 Sol and Gemini 3.1 Pro Preview really cost you.
everyaitoken reads your Codex, Cursor, OpenCode, OpenRouter, and Gemini CLI history on your Mac and prices every request at API rates, with what caching saved or cost. $9 once.
FAQ
Questions
Is Gemini 3.1 Pro Preview cheaper than GPT-5.6 Sol?
Yes, on every workload here. The example agentic session costs $2.00 against $4.20, and a month of 110 sessions $220.00 against $462.00, a $242.00 difference in API-equivalent terms.
Are GPT-5.6 Sol's prices temporary?
OpenAI calls them promotional rates, available at least through November 21, 2026. The sources behind this page do not list a later price.
Where does GPT-5.6 Sol still run in Codex?
Codex cloud chats on ChatGPT plans use it. Outside Codex cloud, Codex suggests moving to GPT-6 Sol.
Which of the two works in GitHub Copilot?
GPT-5.6 Sol is listed in GitHub Copilot. Copilot retired Gemini 3.1 Pro Preview on September 1, 2026, and it remains available in Gemini CLI, Cursor, OpenCode, and OpenRouter.
Sources
- OpenAI: API pricing
- OpenAI docs: GPT-5.6 Sol
- OpenAI: Using GPT-5.6
- OpenAI: API changelog
- Codex docs: Models
- Codex docs: Pricing
- Cursor docs: Models and pricing
- OpenRouter: GPT-5.6 Sol
- OpenCode docs: Zen
- Google: Gemini API pricing
- Google docs: Gemini 3.1 Pro Preview
- Google: Gemini models
- Google: Gemini 3.1 Pro
- Google DeepMind: Gemini
- Gemini CLI source: model configuration
- Cursor docs: Gemini 3.1 Pro
- OpenRouter: Gemini 3.1 Pro Preview
- GitHub Docs: Supported AI models in Copilot
- OpenAI: Prompt caching
- Google: Context caching