Skip to content

Model vs model

GPT-6 Sol vs DeepSeek V4 Pro

GPT-6 Sol and DeepSeek V4 Pro are both in llmwise. What each gets on every plan, what it reads, how it's served and what happens when its provider fails, from the catalog and the code that runs them.

Model prices and specs checked against OpenRouter's GPT-6 Sol page, OpenRouter's DeepSeek V4 Pro page. Updated .

Short answer

DeepSeek V4 Pro gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for GPT-6 Sol. Otherwise, only GPT-6 Sol reads images and only GPT-6 Sol reads a PDF as the whole file. In our test runs, GPT-6 Sol passed 45 of the 50 prompts both answered and DeepSeek V4 Pro 45; 8 prompts split them, most on coding (5 to 4).

GPT-6 Sol vs DeepSeek V4 Pro, prompt by prompt

Every prompt GPT-6 Sol and DeepSeek V4 Pro both answered, compared directly, their biggest differences first. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.

Of the 50 prompts both answered, both passed 41, only GPT-6 Sol passed 4, only DeepSeek V4 Pro passed 4, and neither passed 1. GPT-6 Sol answered sooner on 39 of the 50 and DeepSeek V4 Pro on 10; the rest were within 10% of each other. The 50 replies cost $0.1297 on GPT-6 Sol and $0.1030 on DeepSeek V4 Pro: 1.3× less on DeepSeek V4 Pro.

The 8 prompts only one of GPT-6 Sol and DeepSeek V4 Pro passed

  • Parse CSV with quoted fields (coding): GPT-6 Sol passed and DeepSeek V4 Pro didn't. GPT-6 Sol: All 8 tests passed. DeepSeek V4 Pro: No answer within Pro's reply limit of 8,000 tokens: the model spent them all reasoning.

  • Rewrite corporate jargon in plain words (writing): DeepSeek V4 Pro passed and GPT-6 Sol didn't. GPT-6 Sol: Graded 3.7 of 5 on average (lowest 3). DeepSeek V4 Pro: Graded 4.0 of 5 on average (lowest 3).

  • A product announcement with five rules (writing): GPT-6 Sol passed and DeepSeek V4 Pro didn't. GPT-6 Sol: Graded 4.7 of 5 on average (lowest 4). DeepSeek V4 Pro: Graded 3.7 of 5 on average (lowest 2); but doesn't end with a question.

  • Argue both sides of free buses (writing): GPT-6 Sol passed and DeepSeek V4 Pro didn't. GPT-6 Sol: Graded 4.3 of 5 on average (lowest 4). DeepSeek V4 Pro: Graded 4.0 of 5 on average (lowest 4); but a paragraph of 95 words, over the 90 allowed.

  • An article in three bullets (summarization): DeepSeek V4 Pro passed and GPT-6 Sol didn't. GPT-6 Sol: Graded 4.0 of 5 on average (lowest 3); but 66 words, over the 60 allowed. DeepSeek V4 Pro: Graded 4.3 of 5 on average (lowest 4).

  • An email thread in one sentence (summarization): GPT-6 Sol passed and DeepSeek V4 Pro didn't. GPT-6 Sol: Graded 5.0 of 5 on average (lowest 5). DeepSeek V4 Pro: Graded 4.7 of 5 on average (lowest 4); but 33 words, over the 30 allowed.

  • A late order (customer support): DeepSeek V4 Pro passed and GPT-6 Sol didn't. GPT-6 Sol: Graded 3.7 of 5 on average (lowest 1). DeepSeek V4 Pro: Graded 4.3 of 5 on average (lowest 3).

  • Idioms into natural Japanese (translation): DeepSeek V4 Pro passed and GPT-6 Sol didn't. GPT-6 Sol: Back-translation chrF 0.40 (pass at 0.4); reads back too far from the original. DeepSeek V4 Pro: Back-translation chrF 0.63 (pass at 0.4).

Job by job, the widest gaps first

  • Writing: GPT-6 Sol passed 4 of 5 and DeepSeek V4 Pro 3 of 5. DeepSeek V4 Pro answered 1.5× sooner at the median, 3.5 s against 2.2 s. DeepSeek V4 Pro cost 5.1× less, $0.0113 against $0.0022 for the 5 replies. DeepSeek V4 Pro's replies ran 23% longer, in tokens of reply, thinking not counted.

  • Translation: GPT-6 Sol passed 4 of 5 and DeepSeek V4 Pro 5 of 5. GPT-6 Sol answered 1.3× sooner at the median, 3.2 s against 4.3 s. DeepSeek V4 Pro cost 3.1× less, $0.0140 against $0.0045 for the 5 replies. Their replies ran to about the same length.

  • Coding: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 4 of 5. GPT-6 Sol answered 4.2× sooner at the median, 4.8 s against 20.1 s. GPT-6 Sol cost 2.5× less, $0.0262 against $0.0662 for the 5 replies. DeepSeek V4 Pro's replies ran 28% longer, in tokens of reply, thinking not counted.

  • Customer support: GPT-6 Sol passed 3 of 5 and DeepSeek V4 Pro 4 of 5. GPT-6 Sol answered 1.6× sooner at the median, 3.2 s against 5.1 s. DeepSeek V4 Pro cost 2.4× less, $0.0146 against $0.0061 for the 5 replies. DeepSeek V4 Pro's replies ran 35% longer, in tokens of reply, thinking not counted.

  • Agents and tool use: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 5 of 5. GPT-6 Sol answered 1.8× sooner at the median, 1.9 s against 3.3 s. DeepSeek V4 Pro cost 5.9× less, $0.0081 against $0.0014 for the 5 replies. DeepSeek V4 Pro's replies ran 21% longer, in tokens of reply, thinking not counted.

  • RAG and answering from documents: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 5 of 5. GPT-6 Sol answered 1.7× sooner at the median, 1.1 s against 1.9 s. DeepSeek V4 Pro cost 4.8× less, $0.0081 against $0.0017 for the 5 replies. GPT-6 Sol's replies ran 10% longer, in tokens of reply, thinking not counted.

  • Math: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 5 of 5. GPT-6 Sol answered 1.3× sooner at the median, 2.2 s against 2.7 s. DeepSeek V4 Pro cost 2.7× less, $0.0090 against $0.0033 for the 5 replies. DeepSeek V4 Pro's replies ran 20% longer, in tokens of reply, thinking not counted.

  • Summarization: GPT-6 Sol passed 4 of 5 and DeepSeek V4 Pro 4 of 5. GPT-6 Sol answered 1.5× sooner at the median, 2.0 s against 3.1 s. DeepSeek V4 Pro cost 2.4× less, $0.0110 against $0.0045 for the 5 replies. Their replies ran to about the same length.

  • SQL: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 5 of 5. GPT-6 Sol answered 2.8× sooner at the median, 1.3 s against 3.8 s. DeepSeek V4 Pro cost 2.1× less, $0.0104 against $0.0049 for the 5 replies. Their replies ran to about the same length.

  • Data analysis: GPT-6 Sol passed 5 of 5 and DeepSeek V4 Pro 5 of 5. DeepSeek V4 Pro answered 1.1× sooner at the median, 2.8 s against 2.5 s. DeepSeek V4 Pro cost 2.1× less, $0.0169 against $0.0081 for the 5 replies. DeepSeek V4 Pro's replies ran 40% longer, in tokens of reply, thinking not counted.

All 50 prompts: who passed, who answered sooner, who cost less
GPT-6 Sol and DeepSeek V4 Pro on each prompt of our test runs
PromptResultSoonerCheaper
Turn a title into a URL slugBoth passedDeepSeek V4 Pro, 1.5×took 3.3 s and 2.1 sDeepSeek V4 Pro, 2.2×cost $0.0028 and $0.0013
Parse a duration like “1h 30m”Both passedGPT-6 Sol, 4.2×took 4.8 s and 20.1 sDeepSeek V4 Pro, 1.4×cost $0.0042 and $0.0030
Merge overlapping intervalsBoth passedGPT-6 Sol, 4.0×took 1.7 s and 6.6 sDeepSeek V4 Pro, 4.5×cost $0.0022 and $0.0005
Evaluate an arithmetic expression, no evalBoth passedGPT-6 Sol, 4.1×took 9.7 s and 40.3 sGPT-6 Sol, 3.3×cost $0.0089 and $0.0291
Parse CSV with quoted fieldsOnly GPT-6 SolGPT-6 Sol, 7.8×took 8.9 s and 69.1 sGPT-6 Sol, 4.0×cost $0.0081 and $0.0323
Announce a second bakery shop on LinkedInBoth passedDeepSeek V4 Pro, 1.8×took 3.5 s and 1.9 sDeepSeek V4 Pro, 3.2×cost $0.0027 and $0.0008
Rewrite corporate jargon in plain wordsOnly DeepSeek V4 ProDeepSeek V4 Pro, 4.3×took 5.0 s and 1.2 sDeepSeek V4 Pro, 5.6×cost $0.0033 and $0.0006
Decline a meeting and offer two timesBoth passedGPT-6 Sol, 1.7×took 1.3 s and 2.2 sDeepSeek V4 Pro, 4.2×cost $0.0014 and $0.0003
A product announcement with five rulesOnly GPT-6 SolGPT-6 Sol, 2.4×took 1.6 s and 3.8 sDeepSeek V4 Pro, 10.0×cost $0.0014 and $0.0001
Argue both sides of free busesOnly GPT-6 SolGPT-6 Sol, 1.9×took 4.2 s and 7.8 sDeepSeek V4 Pro, 8.0×cost $0.0025 and $0.0003
A discount, then sales taxBoth passedGPT-6 Sol, 1.6×took 1.7 s and 2.7 sDeepSeek V4 Pro, 6.5×cost $0.0013 and $0.0002
Pens at 3 for $4Both passedDeepSeek V4 Pro, 1.3×took 3.3 s and 2.6 sDeepSeek V4 Pro, 1.6×cost $0.0024 and $0.0016
Compound interest over three yearsBoth passedGPT-6 Sol, 1.1×took 1.4 s and 1.5 sDeepSeek V4 Pro, 4.0×cost $0.0014 and $0.0004
Four-digit numbers whose digits sum to 9Both passedGPT-6 Sol, 1.3×took 2.5 s and 3.2 sDeepSeek V4 Pro, 3.2×cost $0.0019 and $0.0006
The highest of three dice is a 5Both passedGPT-6 Sol, 1.6×took 2.2 s and 3.5 sDeepSeek V4 Pro, 3.3×cost $0.0020 and $0.0006
An article in three bulletsOnly DeepSeek V4 ProGPT-6 Sol, 3.0×took 2.0 s and 6.1 sClosecost $0.0020 and $0.0021
An email thread in one sentenceOnly GPT-6 SolGPT-6 Sol, 1.4×took 2.0 s and 2.8 sDeepSeek V4 Pro, 9.3×cost $0.0014 and $0.0001
Decisions and action items from a meetingBoth passedDeepSeek V4 Pro, 2.3×took 3.5 s and 1.5 sDeepSeek V4 Pro, 7.1×cost $0.0034 and $0.0005
A quarterly memo for the CEOBoth passedGPT-6 Sol, 1.2×took 2.6 s and 3.1 sDeepSeek V4 Pro, 13.6×cost $0.0024 and $0.0002
A study with a negative resultBoth passedGPT-6 Sol, 2.1×took 1.7 s and 3.5 sClosecost $0.0017 and $0.0016
The region with the most revenueBoth passedDeepSeek V4 Pro, 1.4×took 2.8 s and 2.0 sDeepSeek V4 Pro, 1.8×cost $0.0026 and $0.0014
Average order value in AugustBoth passedGPT-6 Sol, 1.3×took 2.8 s and 3.7 sDeepSeek V4 Pro, 1.2×cost $0.0025 and $0.0021
Revenue change from July to AugustBoth passedDeepSeek V4 Pro, 1.1×took 2.8 s and 2.5 sDeepSeek V4 Pro, 1.5×cost $0.0025 and $0.0017
A median, filtered two waysBoth passedDeepSeek V4 Pro, 2.0×took 2.5 s and 1.3 sDeepSeek V4 Pro, 2.8×cost $0.0024 and $0.0008
Correlation between ad spend and sign-upsBoth passedGPT-6 Sol, 6.3×took 6.6 s and 41.2 sDeepSeek V4 Pro, 3.2×cost $0.0069 and $0.0021
A late orderOnly DeepSeek V4 ProGPT-6 Sol, 6.3×took 3.2 s and 20.2 sGPT-6 Sol, 1.7×cost $0.0025 and $0.0041
A return inside the windowBoth passedGPT-6 Sol, 1.6×took 1.5 s and 2.4 sDeepSeek V4 Pro, 1.8×cost $0.0017 and $0.0009
A frustrated customerNeither passedGPT-6 Sol, 1.5×took 5.9 s and 8.6 sDeepSeek V4 Pro, 8.4×cost $0.0043 and $0.0005
A refund request outside the windowBoth passedGPT-6 Sol, 1.3×took 3.0 s and 3.9 sDeepSeek V4 Pro, 9.0×cost $0.0025 and $0.0003
A message with a planted instructionBoth passedGPT-6 Sol, 1.1×took 4.4 s and 5.1 sDeepSeek V4 Pro, 12.0×cost $0.0037 and $0.0003
A delivery message into SpanishBoth passedGPT-6 Sol, 1.3×took 3.2 s and 4.3 sDeepSeek V4 Pro, 2.8×cost $0.0026 and $0.0009
A product description into FrenchBoth passedGPT-6 Sol, 1.5×took 2.1 s and 3.1 sDeepSeek V4 Pro, 8.3×cost $0.0019 and $0.0002
A meeting note into GermanBoth passedGPT-6 Sol, 4.4×took 5.0 s and 22.0 sDeepSeek V4 Pro, 3.5×cost $0.0035 and $0.0010
Idioms into natural JapaneseOnly DeepSeek V4 ProGPT-6 Sol, 2.0×took 2.4 s and 4.7 sDeepSeek V4 Pro, 1.4×cost $0.0021 and $0.0015
A lease clause into Brazilian PortugueseBoth passedDeepSeek V4 Pro, 1.2×took 4.6 s and 3.7 sDeepSeek V4 Pro, 4.8×cost $0.0040 and $0.0008
Customers in one countryBoth passedGPT-6 Sol, 1.8×took 1.1 s and 1.9 sDeepSeek V4 Pro, 7.1×cost $0.0012 and $0.0002
Count orders by statusBoth passedGPT-6 Sol, 2.6×took 1.1 s and 2.7 sDeepSeek V4 Pro, 6.4×cost $0.0012 and $0.0002
Revenue by categoryBoth passedGPT-6 Sol, 2.8×took 1.3 s and 3.8 sDeepSeek V4 Pro, 9.1×cost $0.0018 and $0.0002
Every customer, even those without ordersBoth passedGPT-6 Sol, 3.2×took 1.4 s and 4.5 sGPT-6 Sol, 1.5×cost $0.0018 and $0.0028
Monthly revenue with a running totalBoth passedGPT-6 Sol, 7.7×took 4.5 s and 34.4 sDeepSeek V4 Pro, 2.7×cost $0.0043 and $0.0016
A fact from one sectionBoth passedGPT-6 Sol, 2.0×took 1.1 s and 2.2 sDeepSeek V4 Pro, 6.2×cost $0.0015 and $0.0002
Core hours and start timesBoth passedGPT-6 Sol, 2.3×took 1.1 s and 2.5 sDeepSeek V4 Pro, 2.8×cost $0.0016 and $0.0006
Two sections in one answerBoth passedClosetook 1.2 s and 1.3 sDeepSeek V4 Pro, 5.6×cost $0.0016 and $0.0003
A later amendment changes the answerBoth passedDeepSeek V4 Pro, 1.5×took 2.0 s and 1.3 sDeepSeek V4 Pro, 5.6×cost $0.0020 and $0.0004
A question the handbook doesn't answerBoth passedGPT-6 Sol, 2.1×took 0.9 s and 1.9 sDeepSeek V4 Pro, 6.3×cost $0.0014 and $0.0002
Pick the tool and work out the dateBoth passedGPT-6 Sol, 1.5×took 2.2 s and 3.2 sDeepSeek V4 Pro, 7.9×cost $0.0016 and $0.0002
Convert a currencyBoth passedGPT-6 Sol, 1.9×took 1.8 s and 3.4 sDeepSeek V4 Pro, 11.2×cost $0.0012 and $0.0001
Book a meeting from a sentenceBoth passedGPT-6 Sol, 1.9×took 1.9 s and 3.5 sDeepSeek V4 Pro, 8.3×cost $0.0021 and $0.0003
Search, but don't bookBoth passedGPT-6 Sol, 1.2×took 1.1 s and 1.3 sDeepSeek V4 Pro, 2.3×cost $0.0014 and $0.0006
Two calls with a unit conversionBoth passedGPT-6 Sol, 1.2×took 2.8 s and 3.3 sDeepSeek V4 Pro, 9.7×cost $0.0018 and $0.0002

GPT-6 Sol vs DeepSeek V4 Pro in our test runs

GPT-6 Sol and DeepSeek V4 Pro on the same prompts, job by job: how many replies passed their check.

Based on 100 of our test runs on , through OpenRouter with the app's own prompt and settings. Updated .

GPT-6 Sol and DeepSeek V4 Pro in our test runs, job by job
JobGPT-6 SolDeepSeek V4 Pro
Coding5 of 54 of 5
Writing4 of 53 of 5
Math5 of 55 of 5
Summarization4 of 54 of 5
Data analysis5 of 55 of 5
Customer support3 of 54 of 5
Translation4 of 55 of 5
SQL5 of 55 of 5
RAG and answering from documents5 of 55 of 5
Agents and tool use5 of 55 of 5
All jobs45 of 50 · 1 message each · $0.0026 a reply45 of 50 · 1 message each · $0.0021 a reply

The same prompts, sent to both through llmwise's own pipeline and checked the same way: five for each job. A bold count is the one that passed more. Messages: what each reply counted as on Pro. Cost: what OpenRouter charged us per reply, on average. Every prompt, and how it's scored.

Where they differed most, coding: both replies to a prompt only one of them passed.

A hard prompt: “Parse CSV with quoted fields”, in full.

  • GPT-6 Sol

    Passed: All 8 tests passed.

    ```javascript
    function parseCsv(text) {
      if (text === "") return [];
    
      const rows = [];
      let row = [];
      let field = "";
      let inQuotes = false;
      let atFieldStart = true;
      let endedWithRowBreak = false;
    
      for (let i = 0; i < text.length; i++) {
        const char = text[i];
    
        if (inQuotes) {
          if (char === '"') {
            if (text[i + 1] === '"') {
              field += '"';
              i++;
            } else {
              inQuotes = false;
            }
          } else {
            field += char;
          }
          endedWithRowBreak = false;
        } else if (char === '"' && atFieldStart) {
          inQuotes = true;
          atFieldStart = false;
          endedWithRowBreak = false;
        } else if (char === ",") {
          row.push(field);
          field = "";
          atFieldStart = true;
          endedWithRowBreak = false;
        } else if (char === "\n" || (char === "\r" && text[i + 1] === "\n")) {
          row.push(field);
          rows.push(row);
          row = [];
          field = "";
          atFieldStart = true;
          endedWithRowBreak = true;
          if (char === "\r") i++;
        } else {
          field += char;
          atFieldStart = false;
          endedWithRowBreak = false;
        }
      }
    
      if (!endedWithRowBreak) {
        row.push(field);
        rows.push(row);
      }
    
      return rows;
    }
    ```

    481 tokens in, 712 out (363 of them reasoning) · 8.9 s · $0.0081 · 1 message on Pro · answered by openai/gpt-6-sol via OpenAI ·

  • DeepSeek V4 Pro

    Failed: No answer within Pro's reply limit of 8,000 tokens: the model spent them all reasoning.

    (no reply)

    493 tokens in, 8,000 out (7,805 of them reasoning) · 69.1 s · $0.0323 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via CoreWeave ·

GPT-6 Sol and DeepSeek V4 Pro on every plan

Whether the one-time free trial reaches each model, then each paid plan's messages on it.

Messages on GPT-6 Sol and DeepSeek V4 Pro, plan by plan
PlanPriceGPT-6 SolDeepSeek V4 Pro
Free$0In the one-time trial of 5 messagesIn the one-time trial of 5 messages
Pro$20 a monthUp to 125 a monthUp to 250 a month
Max$50 a monthUp to 400 a monthUp to 800 a month
Ultra$100 a monthUp to 900 a monthUp to 1,800 a month
Studio$200 a monthUp to 2,000 a monthUp to 4,000 a month

Prices don't include tax, which is added where it applies and shown before you pay. A paid plan's month is one allowance shared by every model, so each monthly count is the most you get if all of it goes to that model. It renews each billing period; everyday models refill daily at 00:00 UTC. Long chats count more per reply. How pricing works.

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: 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.

What differs

  • Messages on Pro

    DeepSeek V4 Pro gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for GPT-6 Sol.

  • Context window

    Both take up to 1.05M tokens of context. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.

  • Images and PDFs

    DeepSeek V4 Pro doesn't read images. DeepSeek V4 Pro gets a PDF's text rather than the file itself.

  • On Free

    Both are in the free trial.

  • Where messages go

    GPT-6 Sol: Sent to OpenAI directly. DeepSeek V4 Pro: Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked.

Fact by fact

GPT-6 Sol and DeepSeek V4 Pro, fact by fact
FactGPT-6 SolDeepSeek V4 Pro
Context window1.05M tokens1.05M tokens
Reads imagesYesNo
PDFsWhole fileText only
ReasoningYesYes
API price (September 2026)$2.00 in / $10.00 out per million tokens$0.44 in / $2.90 out per million tokens
A typical message at API prices (4,000 tokens in, 700 out)$0.0150$0.0038
A $10 top-up adds100 messages200 messages
Where a message goesSent to OpenAI directly.Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked.
If the provider failsIf OpenAI fails before the reply starts (an overload, a server error, a dropped connection), llmwise sends the same request to GPT-6 Sol through OpenRouter instead.When one host is down, OpenRouter moves the request to another host that meets the same rules.
Anthropic's safety fallbackDoesn't applyDoesn't apply
API prices are what our model catalog lists (GPT: OpenAI's list price; DeepSeek: the price of the OpenRouter endpoints llmwise uses, not DeepSeek's own API). In llmwise you pay per message, not per token: the counts above are what you get.

GPT-6 Sol or DeepSeek V4 Pro?

From the facts above and our test runs: the rest is how their answers suit your work, which one chat can show you.

  • Pick GPT-6 Sol: it reads images; it reads a PDF as the whole file, charts and scans included.

  • Pick DeepSeek V4 Pro: more messages for the money, up to 250 messages a month on Pro.

Each model's page, the families, and other pairs

GPT-6 Sol vs DeepSeek V4 Pro is one pair of models. The page below covers the whole families.

Questions

Is GPT-6 Sol or DeepSeek V4 Pro cheaper in llmwise?

DeepSeek V4 Pro gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for GPT-6 Sol. Every paid plan's monthly allowance is shared by all models, so each count is the most you get if it all goes to that model.

Can I try GPT-6 Sol and DeepSeek V4 Pro for free?

Yes: both are in the free trial of 5 messages.

Which has the bigger context window, GPT-6 Sol or DeepSeek V4 Pro?

Neither: both take 1.05M tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.

Can I use GPT-6 Sol and DeepSeek V4 Pro in the same chat?

Yes. Pick GPT-6 Sol for one message and DeepSeek V4 Pro for the next; the second sees the whole chat, including the first one's answer.

Which did better in your test runs, GPT-6 Sol or DeepSeek V4 Pro?

On the same 50 prompts, run on September 27, 2026, GPT-6 Sol passed 45 and DeepSeek V4 Pro passed 45. The table on this page has each job, and every reply is published.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, 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.