GPT-6 Sol · Writing

GPT-6 Sol for writing: our test runs

GPT-6 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.

Start free See pricing

Based on 5 of our test runs on September 27, 2026, through OpenRouter with the app's own prompt and settings. Updated October 15, 2026.

Short answer

In our test runs on September 27, 2026, GPT-6 Sol passed 4 of 5 of our writing prompts (both hard ones): tied 4th of 20 models. Each reply counted as 1 message on Pro, where it has 125 a month, and cost $0.0023 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how GPT-6 Sol's reply did.

GPT-6 Sol on each writing prompt
Prompt	Result	Cost	Time
Announce a second bakery shop on LinkedInEveryday	Passed: Graded 4.8 of 5 on average (lowest 4).	$0.0027	3.5 s
Rewrite corporate jargon in plain wordsEveryday	Failed: Graded 3.7 of 5 on average (lowest 3).	$0.0033	5.0 s
Decline a meeting and offer two timesEveryday	Passed: Graded 5.0 of 5 on average (lowest 5).	$0.0014	1.3 s
A product announcement with five rulesHard	Passed: Graded 4.7 of 5 on average (lowest 4).	$0.0014	1.6 s
Argue both sides of free busesHard	Passed: Graded 4.3 of 5 on average (lowest 4).	$0.0025	4.2 s

Every writing reply from GPT-6 Sol

Announce a second bakery shop on LinkedIn

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

Passed: Graded 4.8 of 5 on average (lowest 4).

We’re happy to share that Crumb & Co. is opening a second shop in Leeds on 14 November.

When we opened our first bakery, we hoped to make good bread and bakes for our neighbours. The support we’ve had since then has meant more than we can say. Opening another shop is a big step for our small team, and we’re grateful to everyone who has helped us get here.

If you’re in Leeds on opening day, please stop by and say hello. Coffee is on us all day, and we’d love to welcome you into the new space.

Thank you for making room for Crumb & Co. in your routine. We can’t wait to meet more of our neighbours.

389 tokens in, 189 out (39 of them reasoning) · 3.5 s · $0.0027 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI · September 27, 2026

Rewrite corporate jargon in plain words

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

Failed: Graded 3.7 of 5 on average (lowest 3).

In Q3, teams across the company worked together to put a more customer-focused approach into practice. In the enterprise customer segment, retention measures improved by 12%, while the rate at which customers left fell by 30 basis points (0.3 percentage points). We also adjusted how we connect with customers across channels to make better use of our capacity.

417 tokens in, 246 out (169 of them reasoning) · 5.0 s · $0.0033 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI · September 27, 2026

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: Vendor Demo

Hi Tom,

Thanks for the invitation. I’m presenting to the board on Wednesday at 11:00, so I can’t make the demo then. Would Thursday at 10:00 or Friday at 14:00 work instead?

Best,
[Your Name]
```
395 tokens in, 63 out · 1.3 s · $0.0014 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI · September 27, 2026
A product announcement with five rules
A hard prompt: “A product announcement with five rules”, in full.
Passed: Graded 4.7 of 5 on average (lowest 4).
```
Meet the Nordvik 2 desk lamp, designed to put light where you need it. Choose from five dimming steps, turn the arm through 270 degrees, and charge it via USB-C. Priced at €79, it ships on 1 December. Where will you put yours?
```
406 tokens in, 63 out · 1.6 s · $0.0014 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI · September 27, 2026

Argue both sides of free buses

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

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

A city should make its buses free because a fare can keep people from jobs, classes, and medical appointments. Removing it would make every trip easier, especially for residents with the least money. Boarding would be faster without payment, and some drivers might leave their cars at home, easing traffic and pollution. Funding buses through taxes treats mobility as a public service: everyone benefits when more people can reach what they need.

But free fares are not free service. Fare revenue helps pay for frequent, reliable buses; lose it without secure replacement funding, and riders may face longer waits and crowded routes. Even with new taxes, a city could spend that money on more drivers and better coverage instead. A zero-dollar fare does little for someone whose bus rarely comes. Targeted discounts for low-income riders would protect access while preserving resources to make the service worth taking.

384 tokens in, 173 out · 4.2 s · $0.0025 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI · September 27, 2026

How it compares on writing

GPT-6 Sol beside the job's picks among every model. 3 models share the hard-problems pick, and GPT-6 Sol isn't one of them.

GPT-6 Sol beside other models on writing
Model	Passed	On Pro	Cost per reply
DeepSeek V4 ProOur pick: hard problems (shared), best value	5 of 5	250 a month on Pro	$0.0011
GPT-6 LunaOur pick: hard problems (shared), everyday	5 of 5	60 a day on Pro	$0.0001
GPT-6 Sol	4 of 5	125 a month on Pro	$0.0023

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 Sol next to each model on writing

GPT-6 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 5.1: GPT-6 Sol answered 1.7× sooner (3.5 s to 6.0 s), cost 7.6× less, and passed 4 of 5 to its 3. Only GPT-6 Sol passed A product announcement with five rules; Argue both sides of free buses. Only Claude Fable 5.1 passed Rewrite corporate jargon in plain words.
Against Claude Opus 5.5: GPT-6 Sol answered 2.4× sooner (3.5 s to 8.2 s), cost 8.2× less, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Argue both sides of free buses. Only Claude Opus 5.5 passed Rewrite corporate jargon in plain words.
Against Claude Sonnet 5.5: GPT-6 Sol answered 1.5× later (3.5 s to 2.3 s), cost 1.5× less, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Argue both sides of free buses. Only Claude Sonnet 5.5 passed Rewrite corporate jargon in plain words.
Against Claude Sonnet 5: GPT-6 Sol answered 1.3× sooner (3.5 s to 4.5 s), cost 1.5× less, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Announce a second bakery shop on LinkedIn. Only Claude Sonnet 5 passed Rewrite corporate jargon in plain words.
Against Claude Haiku 5.5: GPT-6 Sol answered 2.3× later (3.5 s to 1.5 s), cost 14.1× more, and passed 4 of 5 to its 3. Only GPT-6 Sol passed Argue both sides of free buses.
Against Claude Haiku 4.5: GPT-6 Sol answered 1.6× later (3.5 s to 2.2 s), cost 2.0× more, and passed 4 of 5 to its 4. Only GPT-6 Sol passed A product announcement with five rules. Only Claude Haiku 4.5 passed Rewrite corporate jargon in plain words.
Against GPT-6 Astra: GPT-6 Sol answered 1.4× sooner (3.5 s to 4.8 s), cost 4.9× less, and passed 4 of 5 to its 5. Only GPT-6 Astra passed Rewrite corporate jargon in plain words.
Against GPT-6.1 Sol: GPT-6 Sol answered within 10% of its time (3.5 s to 3.3 s), cost 2.0× more, and passed 4 of 5 to its 4.
Against GPT-6 Luna: GPT-6 Sol answered 1.4× later (3.5 s to 2.4 s), cost 22.8× more, and passed 4 of 5 to its 5. Only GPT-6 Luna passed Rewrite corporate jargon in plain words.
Against Gemini 4 Argon: GPT-6 Sol answered 1.7× sooner (3.5 s to 5.7 s), cost 1.6× more, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Argue both sides of free buses. Only Gemini 4 Argon passed Rewrite corporate jargon in plain words.
Against Gemini 3.1 Pro (preview): GPT-6 Sol answered 2.6× sooner (3.5 s to 9.1 s), cost 4.1× less, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Argue both sides of free buses. Only Gemini 3.1 Pro (preview) passed Rewrite corporate jargon in plain words.
Against Gemini 3.8 Flash: GPT-6 Sol answered 1.1× sooner (3.5 s to 3.9 s), cost 1.7× more, and passed 4 of 5 to its 3. Only GPT-6 Sol passed A product announcement with five rules.
Against DeepSeek V4.1 Flash: GPT-6 Sol answered 2.6× later (3.5 s to 1.3 s), cost 5.0× more, and passed 4 of 5 to its 4.
Against DeepSeek V4 Pro: GPT-6 Sol answered 1.1× sooner (3.5 s to 3.9 s), cost 2.1× more, and passed 4 of 5 to its 5. Only DeepSeek V4 Pro passed Rewrite corporate jargon in plain words.
Against Grok 4.7: GPT-6 Sol answered 1.7× sooner (3.5 s to 6.0 s), cost 2.5× less, and passed 4 of 5 to its 3. Only GPT-6 Sol passed Argue both sides of free buses.
Against Kimi K3: GPT-6 Sol answered 1.2× later (3.5 s to 3.0 s), cost 1.7× less, and passed 4 of 5 to its 4. Only GPT-6 Sol passed A product announcement with five rules. Only Kimi K3 passed Rewrite corporate jargon in plain words.
Against GLM 5.3: GPT-6 Sol answered 2.0× later (3.5 s to 1.8 s), cost 3.9× more, and passed 4 of 5 to its 4. Only GPT-6 Sol passed Argue both sides of free buses. Only GLM 5.3 passed Rewrite corporate jargon in plain words.
Against GLM 5.3 Flash: GPT-6 Sol answered 2.5× sooner (3.5 s to 8.5 s), cost 7.1× more, and passed 4 of 5 to its 2. Only GPT-6 Sol passed A product announcement with five rules; Argue both sides of free buses.
Against Mistral Large 4: GPT-6 Sol answered 2.8× sooner (3.5 s to 9.8 s), cost within 10% of it, and passed 4 of 5 to its 4. Only GPT-6 Sol passed A product announcement with five rules. Only Mistral Large 4 passed Rewrite corporate jargon in plain words.

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.

Questions

Is GPT-6 Sol good for writing?

In our test runs it passed 4 of 5 writing prompts, tied 4th of the 20 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 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. Free is 5 messages to try; sign in with an email link, no password or card.