Model comparison
Inkling vs Qwen3 32B
Inkling is the stronger model overall, scoring 44.1 to 39.2 on the Noometry Index. Qwen3 32B costs 2.1× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
Alibaba (Qwen)Summary
- They share 21 benchmarks with published results for both. Inkling scores higher in 6 categories and Qwen3 32B in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling leads 40.4 to 20.2.
- The biggest single-benchmark swing is GPQA Diamond: 88.3% for Inkling and 65.7% for Qwen3 32B.
- Qwen3 32B is cheaper at $0.70 / $2.80 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Qwen3 32B accepts more context: 131K tokens versus 66K.
Side by side
| Inkling | Qwen3 32B | |
|---|---|---|
| Provider | Thinking Machines Lab | Alibaba (Qwen) |
| Noometry Index | 44.1 | 39.2 |
| Released | 2026-07-15 | 2025-04 |
| Weights | Open | Open |
| Context window | 66K | 131K |
| Max output | 66K | 16K |
| Input $ / M tokens | $1.87 | $0.70 |
| Output $ / M tokens | $4.68 | $2.80 |
| Results tracked | 41 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 32B leads
Inkling: 34.5 (#234), Qwen3 32B: 37.7 (#190)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| SciCode | 47% | 35.4% |
| LMArena Coding | 1464 | 1358 |
| FrontierCode | 14% | — |
| Aider Polyglot | — | 40% |
| LMArena WebDev | 1413 | — |
| FrontierSWE | 4.1% | — |
| WeirdML | 32.3% | — |
| ALE-Bench | 946 | — |
Agentic & Tool Use Qwen3 32B leads
Inkling: 29.6 (#85), Qwen3 32B: 32.6 (#62)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| APEX-Agents | 33.8% | — |
| Berkeley Function Calling Leaderboard | — | 48.7% |
| τ²-bench Banking | 25% | — |
Reasoning Inkling leads
Inkling: 40.4 (#56), Qwen3 32B: 20.2 (#241)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| CritPt | 5.4% | 0.3% |
| Chess Puzzles | 21% | 5% |
| LMArena Hard Prompts | 1451 | 1334 |
| DTBench | 87.5% | 67.5% |
| LMCA | 37.6% | 17.3% |
| Epoch Capabilities Index | 148.54 | 138.51 |
| ARC-AGI-2 | 36.5% | — |
| SimpleBench | 50% | — |
| Kagi LLM Benchmark | — | 54.9% |
| ARC-AGI-1 | 79.5% | — |
Math Qwen3 32B leads
Inkling: 31.3 (#225), Qwen3 32B: 39.7 (#99)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 66.9% |
| LMArena Math | 1479 | 1399 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| ProofBench | 0% | — |
Knowledge Inkling leads
Inkling: 55.1 (#49), Qwen3 32B: 40.0 (#125)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 88.3% | 65.7% |
| LMArena Expert | 1465 | 1362 |
| SimpleQA Verified | 40.3% | — |
| Vectara Hallucination Rate | — | 5.9% |
Multilingual Inkling leads
Inkling: 54.0 (#52), Qwen3 32B: 45.6 (#167)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1434 | 1317 |
| LMArena Chinese | 1490 | 1357 |
| LMArena German | 1446 | 1341 |
| LMArena Russian | 1429 | 1311 |
| LMArena French | 1458 | — |
| LMArena Japanese | 1429 | — |
| LMArena Korean | 1404 | — |
| LMArena Spanish | 1448 | — |
Instruction Following Inkling leads
Inkling: 75.1 (#71), Qwen3 32B: 68.9 (#179)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1426 | 1305 |
Long Context Too close to call
Inkling: 43.8 (#86), Qwen3 32B: 43.8 (#87)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1434 | 1327 |
| Fiction.LiveBench | — | 74.2% |
Writing & Preference Inkling leads
Inkling: 65.2 (#51), Qwen3 32B: 52.9 (#163)
| Benchmark | Inkling | Qwen3 32B |
|---|---|---|
| LMArena Text | 1441 | 1340 |
| LMArena Creative Writing | 1387 | 1297 |
| LMArena Multi-Turn | 1436 | 1331 |
| EQ-Bench Creative Writing | 1611 | — |
| EQ-Bench 4 | 1226 | — |
Frequently asked questions
Is Inkling better than Qwen3 32B?
Inkling is the stronger model overall, scoring 44.1 to 39.2 on the Noometry Index. Qwen3 32B costs 2.1× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Inkling or Qwen3 32B?
Qwen3 32B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or Qwen3 32B better for coding?
Qwen3 32B scores higher on coding benchmarks: 37.7 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
Qwen3 32B does, with 131K tokens against 66K.
How many benchmarks do Inkling and Qwen3 32B share?
21 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Qwen3 32B has 26.