Model comparison
Inkling vs Qwen2.5-Coder-32B
Inkling is the stronger model overall, scoring 44.1 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 3.5× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 13 shared benchmarks.
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. Inkling scores higher in 7 categories and Qwen2.5-Coder-32B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Inkling leads 65.2 to 41.6.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Inkling accepts more context: 66K tokens versus 33K.
Side by side
| Inkling | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | Thinking Machines Lab | Alibaba (Qwen) |
| Noometry Index | 44.1 | 33.4 |
| Released | 2026-07-15 | 2024-09-18 |
| Weights | Open | Open |
| Context window | 66K | 33K |
| Max output | 66K | 29K |
| Input $ / M tokens | $1.87 | $0.66 |
| Output $ / M tokens | $4.68 | $1 |
| Results tracked | 41 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling leads
Inkling: 34.5 (#234), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Coding | 1464 | 1276 |
| FrontierCode | 14% | — |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| LMArena WebDev | 1413 | — |
| FrontierSWE | 4.1% | — |
| SciCode | 47% | — |
| WeirdML | 32.3% | — |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| ALE-Bench | 946 | — |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Agentic & Tool Use Not comparable
Inkling: 29.6 (#85), Qwen2.5-Coder-32B: —
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| APEX-Agents | 33.8% | — |
| τ²-bench Banking | 25% | — |
Reasoning Inkling leads
Inkling: 40.4 (#56), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1451 | 1251 |
| Epoch Capabilities Index | 148.54 | 119.49 |
| ARC-AGI-2 | 36.5% | — |
| SimpleBench | 50% | — |
| ARC-AGI-1 | 79.5% | — |
| CritPt | 5.4% | — |
| Chess Puzzles | 21% | — |
| LiveBench Reasoning | — | 42.1% |
| DTBench | 87.5% | — |
| LiveBench Data Analysis | — | 49.9% |
| LMCA | 37.6% | — |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |
Math Qwen2.5-Coder-32B leads
Inkling: 31.3 (#225), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1479 | 1251 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| OTIS Mock AIME 2024-2025 | 88.9% | — |
| ProofBench | 0% | — |
| LiveBench Math | — | 46.6% |
| GSM8K | — | 93% |
Knowledge Inkling leads
Inkling: 55.1 (#49), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1465 | 1221 |
| GPQA Diamond | 88.3% | — |
| SimpleQA Verified | 40.3% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |
Multilingual Inkling leads
Inkling: 54.0 (#52), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1434 | 1205 |
| LMArena Chinese | 1490 | 1222 |
| LMArena Russian | 1429 | 1228 |
| LMArena French | 1458 | — |
| LMArena German | 1446 | — |
| LMArena Japanese | 1429 | — |
| LMArena Korean | 1404 | — |
| LMArena Spanish | 1448 | — |
Instruction Following Inkling leads
Inkling: 75.1 (#71), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1426 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
Long Context Inkling leads
Inkling: 43.8 (#86), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1434 | 1251 |
Writing & Preference Inkling leads
Inkling: 65.2 (#51), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | Inkling | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1441 | 1230 |
| LMArena Creative Writing | 1387 | 1174 |
| LMArena Multi-Turn | 1436 | 1222 |
| EQ-Bench Creative Writing | 1611 | — |
| EQ-Bench 4 | 1226 | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Inkling better than Qwen2.5-Coder-32B?
Inkling is the stronger model overall, scoring 44.1 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 3.5× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Inkling or Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or Qwen2.5-Coder-32B better for coding?
Inkling scores higher on coding benchmarks: 34.5 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
Inkling does, with 66K tokens against 33K.
How many benchmarks do Inkling and Qwen2.5-Coder-32B share?
13 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Qwen2.5-Coder-32B has 31.