Model comparison
Claude Opus 4.6 vs Qwen3-4B
Claude Opus 4.6 is the stronger model overall, scoring 58.2 to 31.9 on the Noometry Index.
Last verified . 5 shared benchmarks.
Alibaba (Qwen)Summary
- They share 5 benchmarks with published results for both. Claude Opus 4.6 scores higher in 4 categories and Qwen3-4B in 0 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Opus 4.6 leads 57.8 to 19.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 94.4% for Claude Opus 4.6 and 52.2% for Qwen3-4B.
- Qwen3-4B has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4.6 | Qwen3-4B | |
|---|---|---|
| Provider | Anthropic | Alibaba (Qwen) |
| Noometry Index | 58.2 | 31.9 |
| Released | 2026-02-04 | 2025-04-29 |
| Weights | Proprietary | Open |
| Context window | 1M | — |
| Max output | 128K | — |
| Input $ / M tokens | $5 | — |
| Output $ / M tokens | $25 | — |
| Results tracked | 68 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Claude Opus 4.6: 57.2 (#20), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| SWE-bench Verified | 78.7% | — |
| FrontierCode | 26.6% | — |
| SWE-bench Verified (bash only) | 75.6% | — |
| LMArena WebDev | 1547 | — |
| SWE-bench Multilingual | 72% | — |
| GSO | 41.2% | — |
| WeirdML | 78% | — |
| LMArena Coding | 1536 | — |
| ALE-Bench | 996.5 | — |
| AlgoTune | 1.47 | — |
Agentic & Tool Use Claude Opus 4.6 leads
Claude Opus 4.6: 51.1 (#4), Qwen3-4B: 27.6 (#100)
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| Terminal-Bench | 79.8% | — |
| APEX-Agents | 46.3% | — |
| Berkeley Function Calling Leaderboard | — | 35.7% |
| Remote Labor Index | 4.2% | — |
| τ²-bench Banking | 27.3% | — |
| Cybench | 93% | — |
| DeepResearch Bench | 55.3% | — |
| GBAEval | 44.1% | — |
| LMArena Search | 1253 | — |
| METR Time Horizons | 78.9% | — |
| Vending-Bench 2 | 8,018 | — |
Reasoning Claude Opus 4.6 leads
Claude Opus 4.6: 57.8 (#23), Qwen3-4B: 19.2 (#268)
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| Chess Puzzles | 17% | 4% |
| ARC-AGI-2 | 69.2% | — |
| SimpleBench | 67.6% | — |
| Kagi LLM Benchmark | 83.6% | — |
| NYT Connections (extended) | 92.1% | — |
| ARC-AGI-1 | 94% | — |
| EnigmaEval | 7.6% | — |
| Thematic Generalization | 80.6% | — |
| EBR-Bench | 12.7% | — |
| LMArena Hard Prompts | 1527 | — |
| Mystery Game Puzzles | 25% | — |
| DTBench | 91.2% | — |
| LMCA | 55.8% | — |
| Epoch Capabilities Index | 155.24 | — |
| ForecastBench | 60 | — |
Math Claude Opus 4.6 leads
Claude Opus 4.6: 63.0 (#31), Qwen3-4B: 29.7 (#240)
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| MathArena Final-Answer Competitions | 78.5% | 38.5% |
| OTIS Mock AIME 2024-2025 | 94.4% | 52.2% |
| FrontierMath (Tiers 1-3) | 66% | — |
| FrontierMath Tier 4 | 26.8% | — |
| ProofBench | 50% | — |
| LMArena Math | 1519 | — |
| FrontierMath (Feb 2025 set) | 40.7% | — |
| FrontierMath Tier 4 (v1) | 22.9% | — |
Knowledge Claude Opus 4.6 leads
Claude Opus 4.6: 61.9 (#26), Qwen3-4B: 33.0 (#208)
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| GPQA Diamond | 90.5% | 52.3% |
| Vectara Hallucination Rate | 12.2% | 5.7% |
| Humanity's Last Exam | 34.4% | — |
| SimpleQA Verified | 47% | — |
| LMArena Expert | 1546 | — |
Multimodal Not comparable
Claude Opus 4.6: 37.3 (#74), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| LMArena Vision | 1316 | — |
| Furniture Assembly | 28.3% | — |
| LMArena Document | 1507 | — |
Multilingual Not comparable
Claude Opus 4.6: 57.9 (#6), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| LMArena Non-English | 1489 | — |
| LMArena Chinese | 1551 | — |
| LMArena French | 1513 | — |
| LMArena German | 1502 | — |
| LMArena Japanese | 1484 | — |
| LMArena Korean | 1464 | — |
| LMArena Russian | 1497 | — |
| LMArena Spanish | 1510 | — |
Instruction Following Not comparable
Claude Opus 4.6: 79.5 (#4), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| LMArena Instruction Following | 1523 | — |
Long Context Not comparable
Claude Opus 4.6: 48.1 (#13), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| CL-bench | 20.7% | — |
| CL-bench Life | 17% | — |
| LMArena Longer Query | 1520 | — |
Writing & Preference Not comparable
Claude Opus 4.6: 73.5 (#10), Qwen3-4B: —
| Benchmark | Claude Opus 4.6 | Qwen3-4B |
|---|---|---|
| LMArena Text | 1503 | — |
| LMArena Creative Writing | 1505 | — |
| EQ-Bench Creative Writing | 1809 | — |
| EQ-Bench 4 | 1223 | — |
| LMArena Multi-Turn | 1513 | — |
Frequently asked questions
Is Claude Opus 4.6 better than Qwen3-4B?
Claude Opus 4.6 is the stronger model overall, scoring 58.2 to 31.9 on the Noometry Index.
How many benchmarks do Claude Opus 4.6 and Qwen3-4B share?
5 benchmarks have published results for both models. Claude Opus 4.6 has 68 scored results on Noometry and Qwen3-4B has 6.