Compare / Qwen3 Max vs Claude Sonnet 4
Qwen3 Max vs Claude Sonnet 4
Which is cheaper, which is more capable, and which does better at each kind of work, from 16 results on 3 sources that measured both models.
Qwen3 Max or Claude Sonnet 4? The short answer
- Qwen3 Max is cheaper: $1.56 against $6.00 per million tokens, 3.8 times less.
- Claude Sonnet 4 scores higher on the Artificial Analysis Intelligence Index: 18.9 against 15.6.
- On our business benchmarks, Qwen3 Max ranks higher for sticking to the facts.
- On our business benchmarks, Claude Sonnet 4 ranks higher for agents and tool use, reasoning and knowledge and writing.
At a glance
| Qwen3 Max | Claude Sonnet 4 | |
|---|---|---|
| Price per million tokens | $0.78 in · $3.90 out | $3.00 in · $15.00 out |
| Blended (3 in : 1 out) | $1.56 | $6.00 |
| Cheapest host, blended | – | $6.00 (Amazon Bedrock) |
| Intelligence Index | 15.6 | 18.9 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | – |
| Context window | 262,144 tokens | 200,000 tokens |
| 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 models available today, for each kind of work, from our own benchmarks and licensed sources. The better rank is in green.
| Use case | Qwen3 Max | Claude Sonnet 4 |
|---|---|---|
| Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls | 151 of 161 | 79 of 161 |
| Reasoning and knowledgeHard questions across science, maths and general knowledge | 117 of 195 | 111 of 195 |
| WritingWhat people prefer in blind comparisons, and judged writing quality | 96 of 187 | 68 of 187 |
| Sticking to the factsSummarising without inventing things, and factual answers | 50 of 153 | 88 of 153 |
Every shared result
| Benchmark | Metric | Qwen3 Max | Claude Sonnet 4 |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,419 | 1,387 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,393 | 1,395 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,418 | 1,433 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,413 | 1,415 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,423 | 1,402 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,405 | 1,398 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 15.6 | 18.9 |
| Artificial Analysisreported | GPQA Diamond% of questions | 76.4% | 77.7% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 11.9% | 10.7% |
| Artificial Analysisreported | IFBench% of instructions | 44.1% | 54.7% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 50.0% | 70.3% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 20.5% | 31.1% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 74.3% | 64.6% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 23.0% | 4.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.33 | 0.36 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 38.6 | 63.7 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.
Quick answers
Is Qwen3 Max better than Claude Sonnet 4?
It depends on the task. Among models available today, on our benchmarks and licensed sources, Qwen3 Max ranks higher for sticking to the facts; Claude Sonnet 4 ranks higher for agents and tool use, reasoning and knowledge and writing.
Which is cheaper, Qwen3 Max or Claude Sonnet 4?
Qwen3 Max costs $1.56 per million tokens (three input to one output), against $6.00 for Claude Sonnet 4.
Run Qwen3 Max and Claude Sonnet 4 on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.