Model comparison
Grok 4.3 vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 43.8 on the Noometry Index. Grok 4.3 costs 1.9× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
Alibaba (Qwen)Summary
- They share 33 benchmarks with published results for both. Grok 4.3 scores higher in 0 categories and Qwen3.8 Max in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 46.0.
- The biggest single-benchmark swing is ProofBench: 11% for Grok 4.3 and 58% for Qwen3.8 Max.
- Grok 4.3 is cheaper at $1.25 / $2.50 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
Side by side
| Grok 4.3 | Qwen3.8 Max | |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 43.8 | 56.8 |
| Released | 2026-04-17 | 2026-08-02 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1M |
| Max output | 30K | 131K |
| Input $ / M tokens | $1.25 | $2 |
| Output $ / M tokens | $2.50 | $6 |
| Results tracked | 40 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
Grok 4.3: 41.6 (#121), Qwen3.8 Max: 53.5 (#29)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena WebDev | 1357 | 1674 |
| SciCode | 47.3% | 53.2% |
| LMArena Coding | 1415 | 1502 |
| DeepSWE | — | 57.5% |
| FrontierSWE | — | 17.8% |
| WeirdML | 49.9% | — |
| ALE-Bench | 944.17 | — |
Agentic & Tool Use Qwen3.8 Max leads
Grok 4.3: 27.7 (#99), Qwen3.8 Max: 45.4 (#14)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| GDP.pdf | 8% | 23.2% |
| APEX-Agents | — | 63.3% |
| τ²-bench Banking | — | 55.1% |
| LMArena Search | 1165 | — |
| Vending-Bench 2 | 35.26 | — |
Reasoning Qwen3.8 Max leads
Grok 4.3: 35.9 (#68), Qwen3.8 Max: 54.4 (#26)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| NYT Connections (extended) | 55.2% | 88.3% |
| CritPt | 8% | 20% |
| Chess Puzzles | 25% | 40% |
| LMArena Hard Prompts | 1396 | 1496 |
| DTBench | 90.7% | 92% |
| LMCA | 38.3% | 46.2% |
| Epoch Capabilities Index | 149.16 | 156.41 |
| Mystery Game Puzzles | — | 38% |
| ForecastBench | 60.3 | — |
Math Qwen3.8 Max leads
Grok 4.3: 46.0 (#74), Qwen3.8 Max: 73.2 (#20)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 42.8% | 74.7% |
| FrontierMath Tier 4 | 14.6% | 46.3% |
| OTIS Mock AIME 2024-2025 | 93.3% | 100% |
| ProofBench | 11% | 58% |
| LMArena Math | 1388 | 1499 |
Knowledge Qwen3.8 Max leads
Grok 4.3: 52.5 (#62), Qwen3.8 Max: 61.7 (#27)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 88.8% | 92.7% |
| SimpleQA Verified | 33.2% | 47.3% |
| LMArena Expert | 1385 | 1507 |
Multimodal Qwen3.8 Max leads
Grok 4.3: 31.6 (#104), Qwen3.8 Max: 37.2 (#75)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1229 | 1314 |
| Blueprint-Bench 2 | 0% | — |
| Furniture Assembly | — | 20% |
Multilingual Qwen3.8 Max leads
Grok 4.3: 50.5 (#120), Qwen3.8 Max: 56.7 (#18)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1385 | 1472 |
| LMArena Chinese | 1422 | 1538 |
| LMArena French | 1412 | 1503 |
| LMArena German | 1395 | 1483 |
| LMArena Japanese | 1379 | 1467 |
| LMArena Korean | 1356 | 1461 |
| LMArena Russian | 1399 | 1481 |
| LMArena Spanish | 1398 | 1492 |
Instruction Following Qwen3.8 Max leads
Grok 4.3: 72.1 (#140), Qwen3.8 Max: 77.6 (#17)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1366 | 1479 |
Long Context Qwen3.8 Max leads
Grok 4.3: 42.5 (#123), Qwen3.8 Max: 45.6 (#31)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1393 | 1489 |
Writing & Preference Qwen3.8 Max leads
Grok 4.3: 58.5 (#118), Qwen3.8 Max: 67.1 (#30)
| Benchmark | Grok 4.3 | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1397 | 1483 |
| LMArena Creative Writing | 1380 | 1479 |
| LMArena Multi-Turn | 1406 | 1489 |
| EQ-Bench 4 | 1075 | — |
Frequently asked questions
Is Grok 4.3 better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 43.8 on the Noometry Index. Grok 4.3 costs 1.9× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, Grok 4.3 or Qwen3.8 Max?
Grok 4.3 is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is Grok 4.3 or Qwen3.8 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 41.6 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Grok 4.3 and Qwen3.8 Max share?
33 benchmarks have published results for both models. Grok 4.3 has 40 scored results on Noometry and Qwen3.8 Max has 39.