7 New Models on Ropewalk (Oct 2026): GPT-6.1 Sol Pro and More

By Ropewalk Team. Tested on 2026-10-23: we ran 1 demo image generation on the day these 7 models went live.

On 22 October 2026, Ropewalk added 7 new models: 6 text models with context windows from 192K to 1.05M tokens, plus Bria Image 3.2 for images. This roundup shows what each one does, what it costs to try in coins, and which to open first for which job.

Quick answer

Seven models went live on 2026-10-22: GPT-6.1 Sol Pro, Gemini 4 Argon, Meta Muse Wonder, Meta Muse Spark Open, StepFun Step 5 Preview, Cohere Command A+ and Bria Image 3.2. For a huge document or a hard problem, start with GPT-6.1 Sol Pro or Gemini 4 Argon ( each). For a cheap long-context chat, pick StepFun Step 5 Preview () or Cohere Command A+ (). For commercial-safe images with readable text, use Bria Image 3.2.

Headliners

GPT-6.1 Sol Pro: the Pro tier with a 1.05M-token window

GPT-6.1 Sol Pro is OpenAI's Pro tier of GPT-6.1 Sol, with a context window of 1.05M tokens, which is roughly 700,000 words in one conversation. In Ropewalk chat that means you can paste or attach a long report, a contract bundle or a large codebase excerpt and ask questions about all of it at once. Use it when the answer matters more than the speed: careful analysis, long drafting and multi-step reasoning over a lot of material. A message starts at , so reserve it for the jobs where a stronger model earns its price.

Gemini 4 Argon: Google's reasoning model with multimodal input

Gemini 4 Argon is Google's new reasoning model, with a 1M-token context window and multimodal input, so it accepts images alongside text. It suits problems where you need the model to think a question through: comparing options, debugging a chain of logic, or reading a chart or screenshot and explaining it. Attach an image in Ropewalk chat, add a question, and ask it to reason step by step. A message starts at , the same starting price as GPT-6.1 Sol Pro, which makes the two a natural head-to-head on your own documents.

Bria Image 3.2: licensed-data images with readable text

Bria Image 3.2 is a 4B-parameter text-to-image model trained entirely on licensed data, which makes it a fit for commercial work where the origin of the training set matters. It is also strong at rendering text inside the picture. We ran the prompt below on 2026-10-23 at the default 1:1 setting, and the hand-painted sign came back with "FRESH MODELS DAILY" spelled correctly on the first try, in about 18 seconds. A generation costs . Change the aspect ratio in the settings when you need a banner or a story format.

Text models

Four more text models join the three headliners, and 3 of them offer a 1M-token context window. Each card below shows the live cost.

Meta Muse Wonder

Meta Muse Wonder is the large reasoning model of Meta's Muse family, with a 1M-token context window. It is the most expensive of the seven, starting at , so use it for hard problems where a careful answer is worth it: multi-step logic, planning and dense analysis.

Meta Muse Spark Open

Meta Muse Spark Open is the open-weights variant of Muse Spark, with a 1M-token context window, starting at . It is a sensible everyday long-context model when you want a large window without the price of a reasoning tier.

StepFun Step 5 Preview

StepFun Step 5 Preview is a sparse Mixture-of-Experts model with 27B active and 600B total parameters and a 1M-token context window, strong at software engineering and professional tasks. At it is the cheapest 1M-token model of the batch, which makes it a good first stop for code review and long-file questions.

Cohere Command A+

Cohere Command A+ is a multilingual chat model with a 192K-token context window, enough for roughly 130,000 words, and it is the cheapest of the seven at . Choose it for writing and answering in several languages and for questions about long documents.

Which model for which job

All 7 models in one table, with the starting price in coins. Start with the job, then open the card to see the live cost.

Model	Type	Context	Starting price	Best for
GPT-6.1 Sol Pro	Text	1.05M		deep analysis of long material
Gemini 4 Argon	Text	1M		reasoning, images plus text
Bria Image 3.2	Image	n/a		licensed-data images, readable text
Meta Muse Wonder	Text	1M		hard multi-step reasoning
Meta Muse Spark Open	Text	1M		everyday long-context chat
StepFun Step 5 Preview	Text	1M		coding and long files
Cohere Command A+	Text	192K		multilingual writing, long documents

FAQ

How many new models arrived on 2026-10-22?

Seven: six text models (GPT-6.1 Sol Pro, Gemini 4 Argon, Meta Muse Wonder, Meta Muse Spark Open, StepFun Step 5 Preview and Cohere Command A+) and one image model, Bria Image 3.2.

Which new model is cheapest to try?

Cohere Command A+ starts at , followed by StepFun Step 5 Preview at . The live cost is always on each model card.

Which new models have a 1M-token context?

GPT-6.1 Sol Pro (1.05M), Gemini 4 Argon, Meta Muse Wonder, Meta Muse Spark Open and StepFun Step 5 Preview all offer about 1M tokens. Cohere Command A+ offers 192K.

Can the new models read images?

Gemini 4 Argon accepts multimodal input, so attach an image to your message. Bria Image 3.2 creates images from text prompts.

Earlier releases: Nano Banana 2.1, FLUX 3 and 13 new models, 18 new models from September 2026 and the week 38 roundup. To start with any model above, open Ropewalk and pick it in a new chat.