Model comparison
Grok 4.20 (Non-Reasoning) vs Qwen3.5 397B-A17B
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 46.0 on the Noometry Index.
Last verified . 31 shared benchmarks.
Alibaba (Qwen)Summary
- They share 31 benchmarks with published results for both. Grok 4.20 (Non-Reasoning) scores higher in 7 categories and Qwen3.5 397B-A17B in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.3 to 34.5.
- The biggest single-benchmark swing is NYT Connections (extended): 85.4% for Grok 4.20 (Non-Reasoning) and 58.9% for Qwen3.5 397B-A17B.
- Qwen3.5 397B-A17B is cheaper at $0.60 / $3.60 per million input/output tokens, against $1.25 / $2.50 for Grok 4.20 (Non-Reasoning).
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 262K.
- Qwen3.5 397B-A17B has downloadable open weights; the other is API-only.
Side by side
| Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B | |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 48.6 | 46.0 |
| Released | 2026-02-17 | 2026-02-01 |
| Weights | Proprietary | Open |
| Context window | 1M | 262K |
| Max output | 30K | 66K |
| Input $ / M tokens | $1.25 | $0.60 |
| Output $ / M tokens | $2.50 | $3.60 |
| Results tracked | 46 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Grok 4.20 (Non-Reasoning): 42.1 (#112), Qwen3.5 397B-A17B: 42.0 (#114)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena WebDev | 1375 | 1400 |
| LMArena Coding | 1459 | 1465 |
| WeirdML | 52.3% | — |
| ALE-Bench | 1,150 | — |
Agentic & Tool Use Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 34.4 (#46), Qwen3.5 397B-A17B: 33.3 (#53)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| τ²-bench Banking | 18% | 9.8% |
| Terminal-Bench | 57.3% | — |
| APEX-Agents | — | 24.9% |
| τ²-bench Airline | — | 81.5% |
| τ²-bench Retail | — | 84.4% |
| τ²-bench Telecom | — | 97.8% |
| LMArena Search | 1189 | — |
| Vending-Bench 2 | 4,663 | — |
Reasoning Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 52.3 (#32), Qwen3.5 397B-A17B: 34.5 (#70)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| Kagi LLM Benchmark | 75% | 73.7% |
| NYT Connections (extended) | 85.4% | 58.9% |
| Chess Puzzles | 24% | 13% |
| Thematic Generalization | 63.8% | 65.1% |
| LMArena Hard Prompts | 1451 | 1448 |
| DTBench | 90.1% | 87.5% |
| LMCA | 38.7% | 37.9% |
| Epoch Capabilities Index | 151.98 | 146.65 |
| ARC-AGI-2 | 65.1% | — |
| ARC-AGI-1 | 89.5% | — |
| Mystery Game Puzzles | — | 18% |
| ForecastBench | 61.4 | — |
Math Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 48.2 (#65), Qwen3.5 397B-A17B: 46.1 (#73)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | 31.2% |
| OTIS Mock AIME 2024-2025 | 92.2% | 88.9% |
| LMArena Math | 1455 | 1454 |
| FrontierMath Tier 4 | 17.1% | — |
| ProofBench | 14% | — |
Knowledge Too close to call
Grok 4.20 (Non-Reasoning): 52.8 (#60), Qwen3.5 397B-A17B: 53.3 (#58)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| GPQA Diamond | 89.3% | 86.4% |
| LMArena Expert | 1439 | 1462 |
| SimpleQA Verified | 30.2% | — |
Multimodal Qwen3.5 397B-A17B leads
Grok 4.20 (Non-Reasoning): 33.3 (#98), Qwen3.5 397B-A17B: 40.7 (#44)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Vision | 1263 | 1263 |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1416 | — |
Multilingual Too close to call
Grok 4.20 (Non-Reasoning): 54.5 (#40), Qwen3.5 397B-A17B: 53.7 (#59)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Non-English | 1441 | 1430 |
| LMArena Chinese | 1481 | 1500 |
| LMArena French | 1476 | 1461 |
| LMArena German | 1465 | 1447 |
| LMArena Japanese | 1449 | 1426 |
| LMArena Korean | 1417 | 1384 |
| LMArena Russian | 1458 | 1429 |
| LMArena Spanish | 1443 | 1441 |
Instruction Following Too close to call
Grok 4.20 (Non-Reasoning): 74.8 (#83), Qwen3.5 397B-A17B: 75.0 (#77)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Instruction Following | 1420 | 1424 |
Long Context Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 45.5 (#34), Qwen3.5 397B-A17B: 44.1 (#74)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Longer Query | 1437 | 1442 |
| CL-bench | 22.2% | — |
| CL-bench Life | 11.9% | — |
Writing & Preference Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 65.7 (#44), Qwen3.5 397B-A17B: 62.3 (#79)
| Benchmark | Grok 4.20 (Non-Reasoning) | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Text | 1451 | 1438 |
| LMArena Creative Writing | 1438 | 1401 |
| EQ-Bench Creative Writing | 1574 | 1478 |
| LMArena Multi-Turn | 1456 | 1446 |
Frequently asked questions
Is Grok 4.20 (Non-Reasoning) better than Qwen3.5 397B-A17B?
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 46.0 on the Noometry Index.
Which is cheaper, Grok 4.20 (Non-Reasoning) or Qwen3.5 397B-A17B?
Qwen3.5 397B-A17B is cheaper. It lists at $0.60 per million input tokens and $3.60 per million output tokens; Grok 4.20 (Non-Reasoning) lists at $1.25 and $2.50.
Is Grok 4.20 (Non-Reasoning) or Qwen3.5 397B-A17B better for coding?
They score almost the same on coding (42.1 vs 42.0); test both on your own repository before choosing.
Which has the bigger context window?
Grok 4.20 (Non-Reasoning) does, with 1M tokens against 262K.
How many benchmarks do Grok 4.20 (Non-Reasoning) and Qwen3.5 397B-A17B share?
31 benchmarks have published results for both models. Grok 4.20 (Non-Reasoning) has 46 scored results on Noometry and Qwen3.5 397B-A17B has 36.