Model comparison
Qwen3.5 35B-A3B vs Qwen3 Max
Qwen3 Max is the stronger model overall, scoring 43.5 to 41.7 on the Noometry Index. Qwen3.5 35B-A3B costs 3.5× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Qwen3.5 35B-A3B scores higher in 3 categories and Qwen3 Max in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in coding, where Qwen3 Max leads 42.4 to 33.5.
- The biggest single-benchmark swing is GPQA Diamond: 83.5% for Qwen3.5 35B-A3B and 72.6% for Qwen3 Max.
- Qwen3.5 35B-A3B is cheaper at $0.25 / $2 per million input/output tokens, against $1.20 / $6 for Qwen3 Max.
- Qwen3.5 35B-A3B has downloadable open weights; the other is API-only.
Side by side
| Qwen3.5 35B-A3B | Qwen3 Max | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 41.7 | 43.5 |
| Released | 2026-02-01 | 2025-09-23 |
| Weights | Open | Proprietary |
| Context window | 262K | 262K |
| Max output | 66K | 66K |
| Input $ / M tokens | $0.25 | $1.20 |
| Output $ / M tokens | $2 | $6 |
| Results tracked | 28 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 Max leads
Qwen3.5 35B-A3B: 33.5 (#244), Qwen3 Max: 42.4 (#93)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| LMArena Coding | 1410 | 1456 |
| LMArena WebDev | 1254 | — |
| SciCode | 29.3% | — |
| ALE-Bench | — | 370.45 |
Agentic & Tool Use Not comparable
Qwen3.5 35B-A3B: —, Qwen3 Max: —
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| Vending-Bench 2 | — | 71.56 |
Reasoning Qwen3.5 35B-A3B leads
Qwen3.5 35B-A3B: 23.9 (#158), Qwen3 Max: 22.0 (#186)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| Chess Puzzles | 10% | 4% |
| LMArena Hard Prompts | 1400 | 1448 |
| DTBench | 80% | 82.1% |
| LMCA | 29.5% | 28.3% |
| Epoch Capabilities Index | 142.52 | 142.41 |
| Kagi LLM Benchmark | — | 72.5% |
| NYT Connections (extended) | — | 30.1% |
| CritPt | 0.6% | — |
| Mystery Game Puzzles | — | 5% |
Math Too close to call
Qwen3.5 35B-A3B: 38.4 (#86), Qwen3 Max: 37.6 (#94)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 70% | 73.3% |
| LMArena Math | 1404 | 1447 |
| FrontierMath (Tiers 1-3) | — | 18.9% |
| MathArena Final-Answer Competitions | 56% | — |
| MATH Level 5 | — | 97.1% |
Knowledge Too close to call
Qwen3.5 35B-A3B: 47.7 (#73), Qwen3 Max: 48.1 (#72)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 83.5% | 72.6% |
| LMArena Expert | 1408 | 1457 |
| SimpleQA Verified | — | 48.7% |
| Vectara Hallucination Rate | 10.5% | — |
Multilingual Qwen3 Max leads
Qwen3.5 35B-A3B: 50.2 (#122), Qwen3 Max: 53.9 (#58)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| LMArena Non-English | 1379 | 1429 |
| LMArena Chinese | 1458 | 1479 |
| LMArena French | 1412 | 1450 |
| LMArena German | 1367 | 1463 |
| LMArena Japanese | 1326 | 1399 |
| LMArena Korean | 1354 | 1397 |
| LMArena Russian | 1377 | 1428 |
| LMArena Spanish | 1390 | 1463 |
Instruction Following Qwen3 Max leads
Qwen3.5 35B-A3B: 73.0 (#123), Qwen3 Max: 75.0 (#81)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| LMArena Instruction Following | 1379 | 1420 |
Long Context Too close to call
Qwen3.5 35B-A3B: 42.4 (#122), Qwen3 Max: 41.6 (#129)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| LMArena Longer Query | 1390 | 1439 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 14.5% |
Writing & Preference Qwen3 Max leads
Qwen3.5 35B-A3B: 58.1 (#118), Qwen3 Max: 62.6 (#72)
| Benchmark | Qwen3.5 35B-A3B | Qwen3 Max |
|---|---|---|
| LMArena Text | 1395 | 1439 |
| LMArena Creative Writing | 1347 | 1401 |
| LMArena Multi-Turn | 1391 | 1446 |
Frequently asked questions
Is Qwen3.5 35B-A3B better than Qwen3 Max?
Qwen3 Max is the stronger model overall, scoring 43.5 to 41.7 on the Noometry Index. Qwen3.5 35B-A3B costs 3.5× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Which is cheaper, Qwen3.5 35B-A3B or Qwen3 Max?
Qwen3.5 35B-A3B is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Qwen3 Max lists at $1.20 and $6.
Is Qwen3.5 35B-A3B or Qwen3 Max better for coding?
Qwen3 Max scores higher on coding benchmarks: 42.4 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 262K tokens.
How many benchmarks do Qwen3.5 35B-A3B and Qwen3 Max share?
23 benchmarks have published results for both models. Qwen3.5 35B-A3B has 28 scored results on Noometry and Qwen3 Max has 33.