Model comparison
Qwen3 Max vs Sonar
Qwen3 Max is the stronger model overall, scoring 43.5 to 37.8 on the Noometry Index. Sonar costs 2.4× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Last verified . 0 shared benchmarks.
Alibaba (Qwen)
PerplexitySummary
- The widest gap is in writing & preference, where Qwen3 Max leads 62.6 to 52.6.
- Sonar is cheaper at $1 / $1 per million input/output tokens, against $1.20 / $6 for Qwen3 Max.
- Qwen3 Max accepts more context: 262K tokens versus 128K.
Side by side
| Qwen3 Max | Sonar | |
|---|---|---|
| Provider | Alibaba (Qwen) | Perplexity |
| Noometry Index | 43.5 | 37.8 |
| Released | 2025-09-23 | 2024-01-01 |
| Weights | Proprietary | Proprietary |
| Context window | 262K | 128K |
| Max output | 66K | 4K |
| Input $ / M tokens | $1.20 | $1 |
| Output $ / M tokens | $6 | $1 |
| Results tracked | 33 | 7 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 Max leads
Qwen3 Max: 42.4 (#93), Sonar: 35.1 (#217)
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| LiveBench Coding | — | 35.1% |
| LMArena Coding | 1456 | — |
| ALE-Bench | 370.45 | — |
Agentic & Tool Use Not comparable
Qwen3 Max: —, Sonar: —
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| Vending-Bench 2 | 71.56 | — |
Reasoning Qwen3 Max leads
Qwen3 Max: 22.0 (#186), Sonar: 20.3 (#221)
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| Kagi LLM Benchmark | 72.5% | — |
| NYT Connections (extended) | 30.1% | — |
| Chess Puzzles | 4% | — |
| LiveBench Reasoning | — | 46.3% |
| LMArena Hard Prompts | 1448 | — |
| Mystery Game Puzzles | 5% | — |
| DTBench | 82.1% | — |
| LiveBench Data Analysis | — | 37.9% |
| LMCA | 28.3% | — |
| Epoch Capabilities Index | 142.41 | — |
| LiveBench | — | 46.9% |
Math Qwen3 Max leads
Qwen3 Max: 37.6 (#94), Sonar: 31.3 (#197)
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| FrontierMath (Tiers 1-3) | 18.9% | — |
| OTIS Mock AIME 2024-2025 | 73.3% | — |
| LiveBench Math | — | 41.6% |
| LMArena Math | 1447 | — |
| MATH Level 5 | 97.1% | — |
Knowledge Not comparable
Qwen3 Max: 48.1 (#72), Sonar: —
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| GPQA Diamond | 72.6% | — |
| SimpleQA Verified | 48.7% | — |
| LMArena Expert | 1457 | — |
Multilingual Not comparable
Qwen3 Max: 53.9 (#58), Sonar: —
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| LMArena Non-English | 1429 | — |
| LMArena Chinese | 1479 | — |
| LMArena French | 1450 | — |
| LMArena German | 1463 | — |
| LMArena Japanese | 1399 | — |
| LMArena Korean | 1397 | — |
| LMArena Russian | 1428 | — |
| LMArena Spanish | 1463 | — |
Instruction Following Qwen3 Max leads
Qwen3 Max: 75.0 (#81), Sonar: 71.5 (#146)
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| LiveBench Instruction Following | — | 76.2% |
| LMArena Instruction Following | 1420 | — |
Long Context Not comparable
Qwen3 Max: 41.6 (#129), Sonar: —
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| Fiction.LiveBench | 66.7% | — |
| CL-bench | 14.5% | — |
| LMArena Longer Query | 1439 | — |
Writing & Preference Qwen3 Max leads
Qwen3 Max: 62.6 (#72), Sonar: 52.6 (#166)
| Benchmark | Qwen3 Max | Sonar |
|---|---|---|
| LMArena Text | 1439 | — |
| LMArena Creative Writing | 1401 | — |
| LMArena Multi-Turn | 1446 | — |
| LiveBench Language | — | 44.1% |
Frequently asked questions
Is Qwen3 Max better than Sonar?
Qwen3 Max is the stronger model overall, scoring 43.5 to 37.8 on the Noometry Index. Sonar costs 2.4× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Which is cheaper, Qwen3 Max or Sonar?
Sonar is cheaper. It lists at $1 per million input tokens and $1 per million output tokens; Qwen3 Max lists at $1.20 and $6.
Is Qwen3 Max or Sonar better for coding?
Qwen3 Max scores higher on coding benchmarks: 42.4 versus 35.1 in the Noometry coding category.
Which has the bigger context window?
Qwen3 Max does, with 262K tokens against 128K.
How many benchmarks do Qwen3 Max and Sonar share?
0 benchmarks have published results for both models. Qwen3 Max has 33 scored results on Noometry and Sonar has 7.