AI models 2 min read

GPT-6.1 Sol vs Gemini 3.8 Flash: token rates, input types and free access

OpenAI's GPT-6.1 Sol against Google's Gemini 3.8 Flash: listed token rates, input types, output limits and whether a free tier is offered.

Listed token rates

OpenAI lists GPT-6.1 Sol at USD 2 per million input tokens and USD 10 per million output tokens for standard uncached text with prompts up to 272K input tokens. Google lists Gemini 3.8 Flash on its Standard paid tier at USD 0.75 per million input tokens and USD 3.75 per million output tokens, thinking tokens included, through 31 December 2026. From 1 January 2027 the same page shows USD 1.50 and USD 7.50. A budget that runs into 2027 should therefore use the later Gemini figures. Both rate sets leave out some charges: Google's excludes cache storage, grounding, other service tiers and taxes, and OpenAI's excludes cache writes, tools and tax.

Inputs and limits

OpenAI describes text and image input with text output, and its page shows audio and video as not supported. Google lists text, image, video, audio and PDF input with text output. OpenAI shows a 1,050,000-token context window and a 128,000-token output ceiling. Google shows an input token limit of 1,048,576 and an output limit of 65,536. The Google figure is an input limit rather than a shared context-window measurement, so the two lengths are not directly comparable. OpenAI also says prompts above 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request. The Google facts held here do not state a separate long-prompt rate.

Free access and status

OpenAI's rate-limit table marks the Free tier as not supported, so GPT-6.1 Sol is listed for paid API usage tiers. Google's pricing table has a Free Tier column that shows input, output and context caching as free of charge; the facts held here do not give that tier's rate limits or data terms. Google describes gemini-3.8-flash as a stable model alias with a latest update of September 2026. Neither page confirms access for a particular account.

Who should read which

If you need audio, video or PDF input, only Google's listed input types cover it. If you need the larger listed output ceiling, OpenAI's 128,000 tokens is higher than Google's 65,536. If you want to try a model without paying, only Google's table lists a free tier. VendorClue did not test either model.

VendorClue read the vendor pages for GPT-6.1 Sol on 18 October 2026 and Gemini 3.8 Flash on 18 October 2026. Prices and plan terms can change, so check the current pages before you buy.

Sources

GPT-6.1 Sol: official API documentation · Evidence: documentation · Oct 18, 2026

Gemini 3.8 Flash: official API documentation · Evidence: documentation · Oct 18, 2026

Gemini 3.8 Flash: official API documentation · Evidence: documentation · Oct 18, 2026

Next step

Tools in this guide

All AI models tools
All guides