Compare / Qwen3.8-Max (0902) vs GPT-6.1 Sol
Qwen3.8-Max (0902) vs GPT-6.1 Sol
Which is cheaper, which is more capable, and which does better at each kind of work, from 65 results on 5 sources that measured both models.
AlibabaQwen3.8-Max (0902)
OpenAIGPT-6.1 Sol
Qwen3.8-Max (0902) wins 318 too close to callGPT-6.1 Sol wins 44
Of the 65 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 13 October 2026 (catalogue release 2026-10-13-2e2420e3c5cc). See also Alibaba vs OpenAI, every model from both labs.
Qwen3.8-Max (0902) or GPT-6.1 Sol? The short answer
Pick GPT-6.1 Sol: it scores higher (Spring Prompt overall 79 against 47) and costs less.
- Spring Prompt overall: Qwen3.8-Max (0902) 47, GPT-6.1 Sol 79 (out of 100; how the score works).
- GPT-6.1 Sol is cheaper to run: $0.008 against $0.026 per product listing on CatalogBench (3.1 times less).
- GPT-6.1 Sol scores higher on the Artificial Analysis Intelligence Index: 51.8 against 45.4.
- GPT-6.1 Sol writes faster: 55 tokens a second against 37.
- GPT-6.1 Sol is newer: released 29 Sep 2026, against 3 Aug 2026 for Qwen3.8-Max (0902).
- On our own benchmarks, on the results both models have: GPT-6.1 Sol does better for product listings and decks from an analysis.
- On third-party benchmarks, on the results both models have: GPT-6.1 Sol does better for agents and tool use, reasoning and knowledge and sticking to the facts.
Head to head
47
Spring Prompt overall
79
45.4
Intelligence Index
51.8
$3.00
Price per million tokens, blendedlower is better
$4.00
37
Output speed, tokens a second
55
1,000,000
Context window, tokens
1,050,000
Every fact side by side
| Qwen3.8-Max (0902) | GPT-6.1 Sol | |
|---|---|---|
| Price per million tokens | $2.00 in · $6.00 out | $2.00 in · $10.00 out |
| Blended (3 in : 1 out) | $3.00 | $4.00 |
| Cheapest host, blended | – | $4.00 (Azure) |
| Cost per product listing (CatalogBench) | $0.026 | $0.008 |
| Cached input, per million | – | $0.10 cached input |
| Intelligence Index | 45.4 | 51.8 |
| Spring Prompt overall | 47 | 79 |
| Output speed, tokens a second | 37 | 55 |
| Context window | 1,000,000 tokens | 1,050,000 tokens |
| Released | 3 Aug 2026 | 29 Sep 2026 (newer) |
| Weights | Closed | Closed |
| Developer based in | China | the US |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Product listingsTurning a sparse product feed and photos into listings that can go live
18 of 21
2 of 21
Decks from an analysisTurning a finished analysis into a deck you could present as it is
12 of 20
3 of 20
User surveysPlanning a user survey and reading its results without being misled
not measured
3 of 9
Marketing planningPlanning a year of ad spend without overspending
15 of 20
not measured
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls
7 of 164
3 of 164
Professional workReal tasks from banking, consulting and law, business documents and freelance projects
18 of 51
9 of 51
Reasoning and knowledgeHard questions across science, maths and general knowledge
17 of 198
5 of 198
WritingWhat people prefer in blind comparisons, and judged writing quality
16 of 188
15 of 188
Sticking to the factsSummarising without inventing things, and factual answers
38 of 152
18 of 152
Speed under pressureGood decisions against a real clock (fast chess)
25 of 25
3 of 24
From our research
- AI Checkpoint: Sol’s new fast lane 9 Oct 2026
- Muse Wonder review: good analysis, work that still needs checking 13 Oct 2026
- We Told 11 AI Models What Happened in 2026. Most of Them Thought We Were Making It Up. 12 Oct 2026
- We Tested the Anatomy Claude Code Skill on Everyday Objects and Seven AI Models 12 Oct 2026
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Qwen3.8-Max (0902) | GPT-6.1 Sol |
|---|---|---|
| CatalogBench measured by Spring Prompt · 3 results | ||
| Cost per productUS dollars, lower is better 15 · 9 of 21 measured | $0.0256 | $0.0084 |
| Reliably publish-ready% of products, higher is better 14 · 1 of 21 measured | 28.6% | 75.0% |
| Reliably publish-ready · sales brief% of products, higher is better 19 · 2 of 21 measured | 0.0% | 67.9% |
| DeckBench measured by Spring Prompt · 3 results | ||
| Cost per deckUS dollars, lower is better 11 · 13 of 20 measured | $0.15 | $0.15 |
| Deck ratingrating, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 12 · 3 of 20 measured | 1,076 | 1,411 |
| Accurate decks% of tasks, higher is better 9 · 1 of 20 measured | 50.0% | 100.0% |
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results | ||
| Overall · TextArena rating, higher is better 20 · 18 of 387 measured | 1,482 | 1,483max reasoning |
| Overall · Text factualityArena rating, higher is better 22 · 37 of 163 measured | 1,472 | 1,463max reasoning |
| Business, management and finance · TextArena rating, higher is better 16 · 27 of 380 measured | 1,484 | 1,475max reasoning |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 14 · 6 of 243 measured | 45.4 | 51.8max reasoning |
| Humanity's Last Exam% of questions, higher is better 24 · 8 of 242 measured | 43.1% | 52.9%max reasoning |
| Terminal-Bench 4.0% of tasks, higher is better 12 · 5 of 112 measured | 38.9% | 56.1%max reasoning |
| SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result | ||
| Correct answers% of questions, higher is better 30 · 2 of 76 measured | 47.3%xhigh reasoning | 73.9%max reasoning |
Show all 65 results (52 more)
| Measure | Qwen3.8-Max (0902) | GPT-6.1 Sol |
|---|---|---|
| CatalogBench measured by Spring Prompt · 27 results | ||
| Channel compliance% of products, higher is better 1 · 1 of 21 measured | 100.0% | 100.0% |
| Channel rules broken% of products, lower is better 11 · 1 of 21 measured | 0.6% | 0.0% |
| Channel rules broken · sales brief% of products, lower is better 20 · 1 of 21 measured | 3.0% | 0.0% |
| Claims to check% of products, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Claims to check · sales brief% of products, lower is better 14 · 1 of 21 measured | 1.3% | 0.0% |
| Conflicts caught% of conflicts, higher is better 1 · 1 of 21 measured | 100.0% | 100.0% |
| Content quality% of checks, higher is better 19 · 1 of 21 measured | 94.2% | 98.6% |
| Content quality · sales brief% of checks, higher is better 20 · 2 of 21 measured | 93.6% | 98.8% |
| Decision accuracy% of decisions, higher is better 17 · 1 of 21 measured | 93.3% | 98.7% |
| Failed outputs% of products, lower is better 20 · 1 of 21 measured | 3.6% | 0.0% |
| Failed outputs · sales brief% of products, lower is better 20 · 1 of 21 measured | 3.6% | 0.0% |
| Field accuracy% of missing fields, higher is better 20 · 4 of 21 measured | 90.3% | 94.4% |
| Invented values% of filled values, lower is better 16 · 15 of 21 measured | 2.7% | 2.2% |
| Missing UK information% of products, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Missing UK information · sales brief% of products, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Not findable% of products, lower is better 4 · 9 of 21 measured | 3.6% | 4.8% |
| Not findable · sales brief% of products, lower is better 2 · 11 of 21 measured | 1.8% | 5.4% |
| Publish-ready listings% of products, higher is better 13 · 1 of 21 measured | 50.0% | 78.0% |
| Publish-ready listings · sales brief% of products, higher is better 19 · 3 of 21 measured | 7.7% | 72.0% |
| Unsupported claimsclaims per product, lower is better 11 · 1 of 21 measured | 0.69 | 0.00 |
| Unsupported claims% of products, lower is better 11 · 1 of 21 measured | 26.8% | 0.0% |
| Unsupported claims · sales brief% of products, lower is better 20 · 3 of 21 measured | 84.5% | 3.6% |
| Unsupported claims · sales briefclaims per product, lower is better 20 · 2 of 21 measured | 4.52 | 0.07 |
| Wrong attributes% of products, lower is better 17 · 9 of 21 measured | 21.4% | 16.7% |
| Wrong attributes · sales brief% of products, lower is better 14 · 12 of 21 measured | 19.6% | 19.1% |
| Wrong category or variant% of products, lower is better 4 · 4 of 21 measured | 1.8% | 1.8% |
| Wrong category or variant · sales brief% of products, lower is better 10 · 4 of 21 measured | 2.4% | 1.8% |
| DeckBench measured by Spring Prompt · 13 results | ||
| Caveat dropped% of tasks, lower is better 11 · 1 of 20 measured | 16.7% | 0.0% |
| Clean layout% of tasks, higher is better 6 · 3 of 20 measured | 66.7% | 83.3% |
| Design quality% of the maximum, higher is better 10 · 8 of 20 measured | 67.0% | 71.4% |
| Draft figure quoted% of tasks, lower is better 12 · 1 of 20 measured | 33.3% | 0.0% |
| Findings missing% of tasks, lower is better 1 · 1 of 20 measured | 0.0% | 0.0% |
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 11 · 5 of 20 measured | 61.3% | 73.7% |
| Layout defects% of tasks, lower is better 6 · 3 of 20 measured | 33.3% | 16.7% |
| Misleading metric used% of tasks, lower is better 1 · 1 of 20 measured | 0.0% | 0.0% |
| Presentable decks% of tasks, higher is better 5 · 2 of 20 measured | 33.3% | 50.0% |
| Recommendation late or wrong% of tasks, lower is better 1 · 1 of 20 measured | 0.0% | 0.0% |
| Slides needing work% of tasks, lower is better 10 · 10 of 20 measured | 50.0% | 50.0% |
| Unsupported claims% of tasks, lower is better 11 · 1 of 20 measured | 16.7% | 0.0% |
| Unsupported numbers% of tasks, lower is better 1 · 1 of 20 measured | 0.0% | 0.0% |
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 8 results | ||
| Confirmed task success · AgentIPS effect estimate, higher is better 20 · 2 of 48 measured | 0.05arena agent | 0.15max reasoning, arena agent |
| Creative writing · TextArena rating, higher is better 17 · 21 of 385 measured | 1,468 | 1,460max reasoning |
| Expert prompts · TextArena rating, higher is better 20 · 4 of 338 measured | 1,513 | 1,543max reasoning |
| Instruction following · TextArena rating, higher is better 25 · 10 of 387 measured | 1,474 | 1,490max reasoning |
| Praise over complaint · AgentIPS effect estimate, higher is better 20 · 6 of 48 measured | 0.03arena agent | 0.26max reasoning, arena agent |
| Steerability · AgentIPS effect estimate, higher is better 18 · 3 of 48 measured | 0.03arena agent | 0.11max reasoning, arena agent |
| Tool grounding · AgentIPS effect estimate, higher is better 43 · 1 of 48 measured | -0.01arena agent | 0.00max reasoning, arena agent |
| Writing, literature and language · TextArena rating, higher is better 18 · 17 of 386 measured | 1,472 | 1,472max reasoning |
| Artificial Analysis reported by Artificial Analysis · 4 results | ||
| Long-context reasoning (AA-LCR)% of questions, higher is better 35 · 6 of 233 measured | 80.3% | 84.0%low reasoning |
| Output speedtokens per second, higher is better 71 · 58 of 74 measured | 37.0 | 55.2max reasoning |
| SciCode% of problems, higher is better 36 · 22 of 113 measured | 52.1% | 55.8%high reasoning |
| Time to first answer tokenseconds, lower is better 74 · 36 of 74 measured | 55.8 | 1.69low reasoning |
Compare them with other models
- Alibaba vs OpenAIEvery model from both labs, best against best
- Qwen3.8-Max (0902) vs Claude Opus 5.5Most intelligent at $3–10 per million tokens
- Qwen3.8-Max (0902) vs Claude Haiku 5.5Cheapest with similar intelligence
- GPT-6.1 Sol vs Claude Opus 5.5Most intelligent at $3–10 per million tokens
- GPT-6.1 Sol vs Muse Spark 1.3Cheapest with similar intelligence
- GPT-6.1 Sol vs GPT-6 AstraJust below overall
Quick answers
Is Qwen3.8-Max (0902) better than GPT-6.1 Sol?
Pick GPT-6.1 Sol: it scores higher (Spring Prompt overall 79 against 47) and costs less.
Which is cheaper, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol is cheaper to run: $0.008 against $0.026 per product listing on CatalogBench (3.1 times less).
Which is faster, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol writes about 55 tokens a second on its usual API, against 37 for Qwen3.8-Max (0902) (Artificial Analysis).
Which is better for product listings, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol ranks 2 of 21 models on sale for turning a sparse product feed and photos into listings that can go live, against 18 for Qwen3.8-Max (0902).
Which is better for decks from an analysis, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol ranks 3 of 20 models on sale for turning a finished analysis into a deck you could present as it is, against 12 for Qwen3.8-Max (0902).
Which is better for agents and tool use, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol ranks 3 of 164 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 7 for Qwen3.8-Max (0902).
Which is better for professional work, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol ranks 9 of 51 models on sale for real tasks from banking, consulting and law, business documents and freelance projects, against 18 for Qwen3.8-Max (0902).
Which has the bigger context window, Qwen3.8-Max (0902) or GPT-6.1 Sol?
GPT-6.1 Sol: 1,050,000 tokens, against 1,000,000 for Qwen3.8-Max (0902).
Run Qwen3.8-Max (0902) and GPT-6.1 Sol on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.