Compare / Qwen3.6-Plus vs Claude Sonnet 4.5
Qwen3.6-Plus vs Claude Sonnet 4.5
Which is cheaper, which is more capable, and which does better at each kind of work, from 28 results on 6 sources that measured both models.
Of the 28 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 9 October 2026 (catalogue release 2026-10-09-dd620e778ac5). See also Alibaba vs Anthropic, every model from both labs.
Qwen3.6-Plus or Claude Sonnet 4.5? The short answer
Pick Qwen3.6-Plus: it scores higher (Intelligence Index 27.0 against 20.7) and costs less.
- Qwen3.6-Plus is cheaper: $0.73 against $6.00 per million tokens, 8.2 times less.
- Qwen3.6-Plus scores higher on the Artificial Analysis Intelligence Index: 27.0 against 20.7.
- Qwen3.6-Plus is newer: released 31 Mar 2026, against 29 Sep 2025 for Claude Sonnet 4.5.
- On third-party benchmarks, on the results both models have: Qwen3.6-Plus does better for reasoning and knowledge; Claude Sonnet 4.5 for writing.
Head to head
Every fact side by side
| Qwen3.6-Plus | Claude Sonnet 4.5 | |
|---|---|---|
| Price per million tokens | $0.33 in · $1.95 out | $3.00 in · $15.00 out |
| Blended (3 in : 1 out) | $0.73 | $6.00 |
| Cheapest host, blended | – | $6.00 (Amazon Bedrock) |
| Intelligence Index | 27.0 | 20.7 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | – |
| Context window | 1,000,000 tokens | 1,000,000 tokens |
| Released | 31 Mar 2026 (newer) | 29 Sep 2025 |
| 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.
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 | Qwen3.6-Plus | Claude Sonnet 4.5 |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results | ||
| Overall · TextArena rating, higher is better 76 · 56 of 387 measured | 1,443 | 1,457high-32k reasoning |
| Overall · Text factualityArena rating, higher is better 84 · 61 of 163 measured | 1,438 | 1,449 |
| Business, management and finance · TextArena rating, higher is better 75 · 51 of 380 measured | 1,446 | 1,461 |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 62 · 101 of 243 measured | 27.0 | 20.7thinking reasoning |
| GPQA Diamond% of questions, higher is better 58 · 94 of 231 measured | 88.2% | 83.4%thinking reasoning |
| Humanity's Last Exam% of questions, higher is better 84 · 116 of 242 measured | 27.8% | 17.8%thinking reasoning |
| OpenHands Index reported by OpenHands Index · 3 results | ||
| Average% resolved, higher is better 16 · 15 of 34 measured | 52.9%openhands | 53.0%openhands |
| Front end (SWE-Bench Multimodal)% resolved, higher is better 20 · 11 of 34 measured | 30.9%openhands | 36.8%openhands |
| Greenfield (Commit0)% resolved, higher is better 14 · 25 of 34 measured | 25.0%openhands | 12.5%openhands |
| SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result | ||
| Correct answers% of questions, higher is better 38 · 58 of 76 measured | 44.1% | 30.7%thinking-59k reasoning |
| UGI Leaderboard reported by UGI Leaderboard · 3 results | ||
| Writing score · UGIscore out of 100, higher is better 49 · 23 of 230 measured | 58.2thinking reasoning | 66.9no reasoning |
| Requested-length error · UGI% off the requested word count, lower is better 25 · 15 of 230 measured | 8.0%thinking reasoning | 5.0%no reasoning |
| Style adherence · UGIscore from 0 to 1, higher is better 96 · 8 of 230 measured | 0.35no reasoning | 0.41thinking reasoning |
| Vending-Bench 2 reported by Vending-Bench 2 · 1 result | ||
| Money after a yearUS dollars, higher is better 25 · 34 of 58 measured | $5,115 | $3,839 |
Show all 28 results (14 more)
| Measure | Qwen3.6-Plus | Claude Sonnet 4.5 |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 4 results | ||
| Creative writing · TextArena rating, higher is better 82 · 27 of 385 measured | 1,407 | 1,454 |
| Expert prompts · TextArena rating, higher is better 68 · 43 of 338 measured | 1,476 | 1,498high-32k reasoning |
| Instruction following · TextArena rating, higher is better 74 · 38 of 387 measured | 1,435 | 1,464high-32k reasoning |
| Writing, literature and language · TextArena rating, higher is better 76 · 38 of 386 measured | 1,423 | 1,454 |
| Artificial Analysis reported by Artificial Analysis · 7 results | ||
| Terminal-Bench 2.1% of tasks, higher is better 47 · 53 of 130 measured | 61.4% | 55.8%thinking reasoning |
| τ-bench banking% of tasks, higher is better 48 · 40 of 121 measured | 20.8% | 24.5%thinking reasoning |
| Artificial Analysis Coding Indexindex score, higher is better 50 · 55 of 131 measured | 54.5 | 52.1thinking reasoning |
| IFBench% of instructions, higher is better 29 · 85 of 195 measured | 75.2% | 57.3%thinking reasoning |
| Long-context reasoning (AA-LCR)% of questions, higher is better 54 · 96 of 233 measured | 78.3% | 72.3%thinking reasoning |
| Terminal-Bench Hard% of tasks, higher is better 23 · 47 of 190 measured | 43.9% | 35.6%thinking reasoning |
| τ²-bench telecom% of tasks, higher is better 7 · 73 of 193 measured | 97.7% | 78.1%thinking reasoning |
| OpenHands Index reported by OpenHands Index · 3 results | ||
| Information gathering (GAIA)% resolved, higher is better 11 · 10 of 34 measured | 72.1%openhands | 72.7%openhands |
| Issue resolution (SWE-Bench)% resolved, higher is better 17 · 17 of 34 measured | 74.2%openhands | 74.2%openhands |
| Testing (SWT-Bench)% resolved, higher is better 21 · 16 of 34 measured | 62.1%openhands | 68.8%openhands |
Compare them with other models
Quick answers
Is Qwen3.6-Plus better than Claude Sonnet 4.5?
Pick Qwen3.6-Plus: it scores higher (Intelligence Index 27.0 against 20.7) and costs less.
Which is cheaper, Qwen3.6-Plus or Claude Sonnet 4.5?
Qwen3.6-Plus is cheaper: $0.73 against $6.00 per million tokens, 8.2 times less.
Which is better for agents and tool use, Qwen3.6-Plus or Claude Sonnet 4.5?
Qwen3.6-Plus ranks 63 of 164 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 99 for Claude Sonnet 4.5.
Which is better for reasoning and knowledge, Qwen3.6-Plus or Claude Sonnet 4.5?
Qwen3.6-Plus ranks 67 of 198 models on sale for hard questions across science, maths and general knowledge, against 98 for Claude Sonnet 4.5.
Which is better for writing, Qwen3.6-Plus or Claude Sonnet 4.5?
Claude Sonnet 4.5 ranks 34 of 188 models on sale for what people prefer in blind comparisons, and judged writing quality, against 54 for Qwen3.6-Plus.
Which is better for sticking to the facts, Qwen3.6-Plus or Claude Sonnet 4.5?
Qwen3.6-Plus ranks 80 of 152 models on sale for summarising without inventing things, and factual answers, against 101 for Claude Sonnet 4.5.
Run Qwen3.6-Plus and Claude Sonnet 4.5 on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.