Skip to content

GPT-6.2 Sol · Writing

GPT-6.2 Sol for writing: our test runs

GPT-6.2 Sol ran our 5 writing prompts in llmwise, through the same pipeline your messages take. Here's every reply as it came, whether it passed, what it cost and what it counts as on Pro, next to the other models.

Based on 5 of our test runs on , through OpenRouter with the app's own prompt and settings.

Short answer

In our test runs on November 17, 2026, GPT-6.2 Sol passed 5 of 5 of our writing prompts (both hard ones): tied 1st of 27 models. Each reply counted as 1 message on Pro, where it has 125 a month, and cost $0.0022 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how GPT-6.2 Sol's reply did.

GPT-6.2 Sol on each writing prompt
PromptResultCostTime
Announce a second bakery shop on LinkedInEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00262.7 s
Rewrite corporate jargon in plain wordsEverydayPassed: Graded 4.3 of 5 on average (lowest 4).$0.00191.8 s
Decline a meeting and offer two timesEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00181.8 s
A product announcement with five rulesHardPassed: Graded 4.0 of 5 on average (lowest 3).$0.00181.8 s
Argue both sides of free busesHardPassed: Graded 5.0 of 5 on average (lowest 5).$0.00304.3 s

Every writing reply from GPT-6.2 Sol

  1. Announce a second bakery shop on LinkedIn

    An everyday prompt: “Announce a second bakery shop on LinkedIn”, in full.

    Passed: Graded 5.0 of 5 on average (lowest 5).

    We’re opening our second Crumb & Co. bakery in Leeds on 14 November, and we’d love to see you there.
    
    Opening another shop is a big step for our small team. It’s one we get to take because so many of you have made us part of your day, whether you stop in for a loaf, pick up a pastry on the way to work, or bring a friend along for coffee. Thank you for coming back and telling people about us.
    
    To celebrate, we’ll be serving free coffee on opening day. If you’re nearby on 14 November, pop in for a cup and say hello. We’re looking forward to meeting our new neighbours and seeing some familiar faces, too.

    563 tokens in, 148 out · 2.7 s · $0.0026 · 1 message on Pro · answered by openai/gpt-6.2-sol via OpenAI ·

  2. Rewrite corporate jargon in plain words

    An everyday prompt: “Rewrite corporate jargon in plain words”, in full.

    Passed: Graded 4.3 of 5 on average (lowest 4).

    In the third quarter, teams across the company worked together to put customers at the center of how we operate. Among enterprise customers, retention measures improved by 12%, and the rate at which customers left fell by 0.30 percentage points. We also adjusted how we interact with customers across channels to make better use of available resources.

    591 tokens in, 71 out · 1.8 s · $0.0019 · 1 message on Pro · answered by openai/gpt-6.2-sol via OpenAI ·

  3. Decline a meeting and offer two times

    An everyday prompt: “Decline a meeting and offer two times”, in full.

    Passed: Graded 5.0 of 5 on average (lowest 5).

    Subject: Rescheduling the vendor demo
    
    Hi Tom,
    
    Thanks for the invitation. I’m presenting to the board on Wednesday at 11:00, so I won’t be able to make the demo. Would Thursday at 10:00 or Friday at 14:00 work for you instead?
    
    Best,  
    [Your Name]

    569 tokens in, 71 out · 1.8 s · $0.0018 · 1 message on Pro · answered by openai/gpt-6.2-sol via OpenAI ·

  4. A product announcement with five rules

    A hard prompt: “A product announcement with five rules”, in full.

    Passed: Graded 4.0 of 5 on average (lowest 3).

    Meet the Nordvik 2 desk lamp. It dims in five steps, and its arm turns 270 degrees so you can direct the light where you need it. It charges via USB-C, costs €79, and ships on 1 December.
    
    Where would you put yours?

    580 tokens in, 60 out · 1.8 s · $0.0018 · 1 message on Pro · answered by openai/gpt-6.2-sol via OpenAI ·

  5. Argue both sides of free buses

    A hard prompt: “Argue both sides of free buses”, in full.

    Passed: Graded 5.0 of 5 on average (lowest 5).

    A city should make its buses free because a fare that looks small on paper can determine whether someone takes a shift, keeps a medical appointment, or buys groceries. Eliminating fares gives riders immediate relief without applications or eligibility checks, and it lets buses spend less time at stops collecting payments. We already fund streets because everyone benefits from a connected city. Free buses apply the same principle to people who can’t afford to drive or would rather not.
    
    A bus that never comes is no bargain, even at zero dollars. Unless the city can reliably replace fare revenue, going fare-free can force cuts that leave riders waiting longer or stranded after a late shift. Even with replacement funding, every dollar spent making trips free is a dollar the city could spend running buses more often or extending routes to underserved neighborhoods. Keep fares affordable, waive them for riders who can’t pay, and put the rest of the money into service people can count on.

    558 tokens in, 189 out · 4.3 s · $0.0030 · 1 message on Pro · answered by openai/gpt-6.2-sol via OpenAI ·

How it compares on writing

GPT-6.2 Sol beside the job's picks among every model. 14 models share the hard-problems pick, GPT-6.2 Sol among them.

GPT-6.2 Sol beside other models on writing
ModelPassedOn ProCost per reply
Gemini 3.8 FlashOur pick: hard problems (shared), best value5 of 5250 a month on Pro$0.0005
DeepSeek V4.1 FlashOur pick: everyday4 of 560 a day on Pro$0.0003
GPT-6.2 SolOur pick: hard problems (shared)5 of 5125 a month on Pro$0.0022
On Pro: Pro's count on each model. Cost: what OpenRouter charged us per reply, on average.

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.

GPT-6.2 Sol next to each model on writing

GPT-6.2 Sol against each model on the same 5 writing prompts. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.

  • Against Claude Fable 5.2: GPT-6.2 Sol answered 1.9× sooner (1.8 s to 3.5 s), cost 5.2× less, and passed 5 of 5 to its 5.

  • Against Claude Fable 5.1: GPT-6.2 Sol answered 2.3× sooner (1.8 s to 4.1 s), cost 6.0× less, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Argue both sides of free buses.

  • Against Claude Opus 5.5: GPT-6.2 Sol answered 2.7× sooner (1.8 s to 4.8 s), cost 7.2× less, and passed 5 of 5 to its 5.

  • Against Claude Sonnet 5.5: GPT-6.2 Sol answered 1.5× sooner (1.8 s to 2.7 s), cost 1.2× less, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Argue both sides of free buses.

  • Against Claude Sonnet 5: GPT-6.2 Sol answered 2.0× sooner (1.8 s to 3.6 s), cost 1.7× less, and passed 5 of 5 to its 3. Only GPT-6.2 Sol passed Announce a second bakery shop on LinkedIn; A product announcement with five rules.

  • Against Claude Haiku 5.5: GPT-6.2 Sol answered within 10% of its time (1.8 s to 1.7 s), cost 16.7× more, and passed 5 of 5 to its 3. Only GPT-6.2 Sol passed Rewrite corporate jargon in plain words; Argue both sides of free buses.

  • Against Claude Haiku 4.5: GPT-6.2 Sol answered within 10% of its time (1.8 s to 1.7 s), cost 1.7× more, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed A product announcement with five rules.

  • Against GPT-6.2 Astra: GPT-6.2 Sol answered 2.7× sooner (1.8 s to 4.9 s), cost 3.9× less, and passed 5 of 5 to its 5.

  • Against GPT-6 Astra: GPT-6.2 Sol answered 2.6× sooner (1.8 s to 4.7 s), cost 2.8× less, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Rewrite corporate jargon in plain words.

  • Against GPT-6.1 Sol: GPT-6.2 Sol answered 3.0× sooner (1.8 s to 5.5 s), cost 1.2× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Sol: GPT-6.2 Sol answered 1.8× sooner (1.8 s to 3.3 s), cost 1.7× more, and passed 5 of 5 to its 5.

  • Against GPT-6.1 Terra: GPT-6.2 Sol answered 2.1× sooner (1.8 s to 3.9 s), cost 3.4× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Luna: GPT-6.2 Sol answered 1.4× sooner (1.8 s to 2.6 s), cost 39.7× more, and passed 5 of 5 to its 3. Only GPT-6.2 Sol passed Rewrite corporate jargon in plain words; Argue both sides of free buses.

  • Against Gemini 4 Carbon: GPT-6.2 Sol answered 11.5× sooner (1.8 s to 21.0 s), cost 2.8× less, and passed 5 of 5 to its 5.

  • Against Gemini 4 Argon: GPT-6.2 Sol answered 4.3× sooner (1.8 s to 7.9 s), cost 1.6× more, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Argue both sides of free buses.

  • Against Gemini 4 Hydrogen: GPT-6.2 Sol answered 4.3× sooner (1.8 s to 7.9 s), cost 3.0× less, and passed 5 of 5 to its 5.

  • Against Gemini 3.1 Pro (preview): GPT-6.2 Sol answered 4.1× sooner (1.8 s to 7.4 s), cost 2.3× less, and passed 5 of 5 to its 3. Only GPT-6.2 Sol passed A product announcement with five rules; Argue both sides of free buses.

  • Against Gemini 3.8 Flash: GPT-6.2 Sol answered 1.7× sooner (1.8 s to 3.1 s), cost 4.5× more, and passed 5 of 5 to its 5.

  • Against DeepSeek V4.1 Flash: GPT-6.2 Sol answered 1.2× later (1.8 s to 1.6 s), cost 6.8× more, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Rewrite corporate jargon in plain words.

  • Against DeepSeek V4 Pro: GPT-6.2 Sol answered 1.8× sooner (1.8 s to 3.3 s), cost 1.7× more, and passed 5 of 5 to its 5.

  • Against Grok 4.7: GPT-6.2 Sol answered 4.7× sooner (1.8 s to 8.6 s), cost 2.5× less, and passed 5 of 5 to its 5.

  • Against Kimi K3.1 Pro: GPT-6.2 Sol answered 3.8× sooner (1.8 s to 6.9 s), cost 2.2× less, and passed 5 of 5 to its 5.

  • Against Kimi K3: GPT-6.2 Sol answered 2.8× sooner (1.8 s to 5.1 s), cost 1.5× less, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Argue both sides of free buses.

  • Against GLM 5.3: GPT-6.2 Sol answered 1.2× later (1.8 s to 1.5 s), cost 2.6× more, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed Argue both sides of free buses.

  • Against GLM 5.3 Flash: GPT-6.2 Sol answered 1.8× later (1.8 s to 1.0 s), cost 13.6× more, and passed 5 of 5 to its 4. Only GPT-6.2 Sol passed A product announcement with five rules.

  • Against Mistral Large 4: GPT-6.2 Sol answered 4.0× sooner (1.8 s to 7.2 s), cost 1.5× more, and passed 5 of 5 to its 5.

How these runs were done

Rubric (graded). The grader model scores the reply from 1 to 5 on each published criterion. It passes with an average of 4 or more and no criterion under 3, and only if it also meets the prompt's automatic rules (length, words it must or mustn't use).

The grader is Claude Opus 5.5 at low reasoning effort; its own replies are graded by GPT-6 Astra, so no model grades itself. Its prompt and every rubric are on the methods page.

How the runs were done, and every writing prompt.

GPT-6.2 Sol, and writing, elsewhere

Questions

Is GPT-6.2 Sol good for writing?

In our test runs it passed 5 of 5 writing prompts, tied 1st of the 27 models in llmwise. Every reply is on this page, so you can judge them yourself.

How many of my messages does a writing reply from GPT-6.2 Sol use?

1 message each on Pro, where it has 125 a month on Pro. The price of a message is fixed and shown before you send it, however long the reply.

How were these runs done?

The same way for every model: each prompt sent through llmwise's own pipeline, each reply checked the same way. The methods page has every prompt and how each is scored.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.

See what a message costs before you send it. Every plan is month to month: cancel in two clicks. Sign in with an email link, no password.