GPT-6.1 Sol vs Qwen3.8 Max
GPT-6.1 Sol scores higher on the Models at Work leaderboard (70 vs 63 out of 100). On independent tests alone they are also level (81 vs 78). Qwen3.8 Max is cheaper: 1,000 support replies cost about $8.40 at list price, against $10 on GPT-6.1 Sol, which is 1.2 times as much. On the Artificial Analysis Intelligence Index, one of the independent tests, GPT-6.1 Sol leads 55 to 34.
| GPT-6.1 Sol · OpenAI | Qwen3.8 Max · Alibaba | |
|---|---|---|
| Leaderboard rank | #6 | #16 |
| Score (0–100) | 70 | 63 |
| Kind of work | Everyday default | Everyday default |
| Price per 1M tokens | $2 to read · $10 to write | $2 to read · $6 to write |
| Reads at once | 1M tokens (about 750,000 words) | 984K tokens (about 738,000 words) |
| Where you can get it | OpenAI API, Amazon Bedrock, Microsoft Foundry, Gemini Enterprise Agent Platform (Google Cloud), ChatGPT Plus, Pro, Business, Enterprise and Edu, Microsoft 365 Copilot | Alibaba Cloud Model Studio |
| Can a company host it itself | No | No |
| Artificial Analysis Intelligence Index | 55 | 34 |
| Humanity's Sixth Sense (Scale) | 46.60 | 33.00 |
| Arena Text score | Not measured | 1483 (rank 22) |
| Arena WebDev score (qwen3.8-max-0902) | Not measured | 1674 (rank 10) |
| Scale MCP Atlas (Qwen3.8-2.4T-A95B open-weight variant, xHigh) | Not measured | 84.50 |
| Best for | Drafting proposals and reports, Summarising long documents, Back-office tasks with several steps, Answering questions from company documents | Multilingual customer chat, Automating multi-step tasks across business systems, Teams already buying from Alibaba Cloud |
Cost of 1,000 jobs
| 1,000 of these | GPT-6.1 Sol | Qwen3.8 Max |
|---|---|---|
| Support repliesthe customer message, the account notes and two help articles read; a 300-word reply written | $10 | $8.40 |
| Documents processeda two-page invoice or form read; the fields pulled out | $7 | $5.80 |
| Calls summariseda 40-minute call transcript read; a summary and next steps written | $26 | $24 |
| Contracts revieweda 40-page contract read against a checklist; a memo of issues written | $75 | $69 |
| Drafts writtena brief and one example read; a 600-word email or proposal section written | $11 | $7.80 |
Worked out from each vendor’s list price and a fixed size for each job. A product built on a model charges its own price on top. See what every model costs for the same jobs.
Choose GPT-6.1 Sol for
- Drafting proposals and reports
- Summarising long documents
- Back-office tasks with several steps
- Answering questions from company documents
The sensible OpenAI default: 55 on Artificial Analysis's index against 57 for GPT-6 Astra, at a fifth of Astra's list price.
Caveats: Two points separate it from Astra on the index, which is within noise; test Astra only where Sol's answers fall short. The ChatGPT plans are listed from our researcher's reading of OpenAI's launch page, which blocked automated reading this run; on ChatGPT Enterprise and Edu an owner has to switch the model on. In Microsoft 365 Copilot it is rolling out in phases across Word, Excel, PowerPoint and Chat with usage limits Microsoft has not published, per Microsoft's 30 September post; admins commenting there report users already reaching them.
Choose Qwen3.8 Max for
- Multilingual customer chat
- Automating multi-step tasks across business systems
- Teams already buying from Alibaba Cloud
Alibaba's hosted flagship scores 34 on Artificial Analysis's index, a little above the open models GLM-5.3 (32) and MiMo-V2.6-Pro (30), at about 1.4 times GLM's list price and about five times MiMo's.
Caveats: The hosted model is closed; the open version is a different model that Artificial Analysis has not scored and that needs data-centre hardware. Slow: about 35 tokens (word-pieces) a second in Artificial Analysis's measurement. The price is Artificial Analysis's listing, not read from Alibaba's own page. Alibaba Cloud Model Studio is listed from our researcher's reading of Alibaba's 3 August announcement; the page could not be re-read this run. It ranks 16th partly because the scoring marks every model on where a business can buy it, and it has one listed place.
What people running them say
GPT-6.1 Sol: The launch thread on Hacker News, a developer forum, argued over its own title: 'near-Astra for a fifth of the price' drew 955 comments, most of them about which work to send to Sol and which to a dearer model.
Qwen3.8 Max: The launch thread on Hacker News, a developer forum (1,124 upvotes, 612 comments), was dominated by people who already run the smaller Qwen 3.6 and 3.8 models on their own computers and are cancelling Claude subscriptions; the Max itself was praised as the first model of Qwen's top class that anyone can download and host. Sceptics (Aurornis) say Qwen output 'has to be discarded for anything other than really easy tasks', and several note that at current electricity prices hosting it yourself costs more than DeepSeek's discounted rate for repeated text.
Questions people ask
Which is better, GPT-6.1 Sol or Qwen3.8 Max?
GPT-6.1 Sol scores higher on the Models at Work leaderboard (70 vs 63 out of 100). On independent tests alone they are also level (81 vs 78). Qwen3.8 Max is cheaper: 1,000 support replies cost about $8.40 at list price, against $10 on GPT-6.1 Sol, which is 1.2 times as much. On the Artificial Analysis Intelligence Index, one of the independent tests, GPT-6.1 Sol leads 55 to 34.
When should I choose GPT-6.1 Sol over Qwen3.8 Max?
Choose GPT-6.1 Sol for drafting proposals and reports, summarising long documents, back-office tasks with several steps, answering questions from company documents. The sensible OpenAI default: 55 on Artificial Analysis's index against 57 for GPT-6 Astra, at a fifth of Astra's list price.
When should I choose Qwen3.8 Max over GPT-6.1 Sol?
Choose Qwen3.8 Max for multilingual customer chat, automating multi-step tasks across business systems, teams already buying from alibaba cloud. Alibaba's hosted flagship scores 34 on Artificial Analysis's index, a little above the open models GLM-5.3 (32) and MiMo-V2.6-Pro (30), at about 1.4 times GLM's list price and about five times MiMo's.
Which is cheaper, GPT-6.1 Sol or Qwen3.8 Max?
GPT-6.1 Sol costs $2 per million tokens it reads (input) and $10 per million tokens it writes (output) on OpenAI's list price (OpenAI API pricing, read Nov 12, 2026). A token is about three-quarters of a word. Qwen3.8 Max costs $2 per million tokens it reads (input) and $6 per million tokens it writes (output) on Alibaba's list price (Qwen3.8 Max (0902) - Artificial Analysis, read Nov 12, 2026). A token is about three-quarters of a word.
What do 1,000 jobs cost on GPT-6.1 Sol and Qwen3.8 Max?
At list price: 1,000 support replies cost $10 on GPT-6.1 Sol and $8.40 on Qwen3.8 Max; 1,000 documents processed cost $7 on GPT-6.1 Sol and $5.80 on Qwen3.8 Max; 1,000 calls summarised cost $26 on GPT-6.1 Sol and $24 on Qwen3.8 Max; 1,000 contracts reviewed cost $75 on GPT-6.1 Sol and $69 on Qwen3.8 Max; 1,000 drafts written cost $11 on GPT-6.1 Sol and $7.80 on Qwen3.8 Max. A product built on either model charges its own price on top.
Scores come from the Models at Work leaderboard: Score is 0 to 100 and computed, not typed. 40% is independent testing (Artificial Analysis, Arena, Scale, Epoch, Vals and others; each test is scaled against the best score on this board, and tests of real work, such as finished freelance projects or professional tasks, count double the exams). 30% is price: the vendor's list price for a typical mix of reading and writing, on a fixed scale from $0.05 to $60 per million tokens, so a model that costs a fifth as much gains real ground. 15% is what people report after running it, from Signals and linked threads. 15% is fit: how much text it can take in at once and where a business can get it (the products and clouds that carry it), with a small credit for open weights. A missing component drops out and the rest are reweighted. A model with no independent test yet is provisional and ranks below every measured one. Within three points is a tie. Vendor-published figures are labelled and never counted. Prices are vendor list prices where published; a typical mix assumes three parts reading to one part writing.