Model comparison
Inkling vs Qwen3.5 27B
Inkling is the stronger model overall, scoring 44.1 to 41.9 on the Noometry Index. Qwen3.5 27B costs 3.1× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. Inkling scores higher in 6 categories and Qwen3.5 27B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Inkling leads 55.1 to 38.0.
- The biggest single-benchmark swing is WeirdML: 32.3% for Inkling and 39.5% for Qwen3.5 27B.
- Qwen3.5 27B is cheaper at $0.30 / $2.40 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Qwen3.5 27B accepts more context: 262K tokens versus 66K.
Side by side
| Inkling | Qwen3.5 27B | |
|---|---|---|
| Provider | Thinking Machines Lab | Alibaba (Qwen) |
| Noometry Index | 44.1 | 41.9 |
| Released | 2026-07-15 | 2026-02-23 |
| Weights | Open | Open |
| Context window | 66K | 262K |
| Max output | 66K | 66K |
| Input $ / M tokens | $1.87 | $0.30 |
| Output $ / M tokens | $4.68 | $2.40 |
| Results tracked | 41 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 27B leads
Inkling: 34.5 (#234), Qwen3.5 27B: 38.9 (#168)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena WebDev | 1413 | 1358 |
| WeirdML | 32.3% | 39.5% |
| LMArena Coding | 1464 | 1427 |
| ALE-Bench | 946 | 349.45 |
| FrontierCode | 14% | — |
| FrontierSWE | 4.1% | — |
| SciCode | 47% | — |
Agentic & Tool Use Not comparable
Inkling: 29.6 (#85), Qwen3.5 27B: —
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| APEX-Agents | 33.8% | — |
| τ²-bench Banking | 25% | — |
| Vending-Bench 2 | — | 201.98 |
Reasoning Inkling leads
Inkling: 40.4 (#56), Qwen3.5 27B: 27.5 (#117)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Hard Prompts | 1451 | 1414 |
| DTBench | 87.5% | 82.4% |
| LMCA | 37.6% | 34% |
| ARC-AGI-2 | 36.5% | — |
| SimpleBench | 50% | — |
| NYT Connections (extended) | — | 47.9% |
| ARC-AGI-1 | 79.5% | — |
| CritPt | 5.4% | — |
| Chess Puzzles | 21% | — |
| Thematic Generalization | — | 45.5% |
| Epoch Capabilities Index | 148.54 | — |
Math Qwen3.5 27B leads
Inkling: 31.3 (#225), Qwen3.5 27B: 38.8 (#127)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Math | 1479 | 1429 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| MathArena Final-Answer Competitions | — | 56.7% |
| OTIS Mock AIME 2024-2025 | 88.9% | — |
| ProofBench | 0% | — |
Knowledge Inkling leads
Inkling: 55.1 (#49), Qwen3.5 27B: 38.0 (#150)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Expert | 1465 | 1428 |
| GPQA Diamond | 88.3% | — |
| SimpleQA Verified | 40.3% | — |
| Vectara Hallucination Rate | — | 12.1% |
Multimodal Not comparable
Inkling: —, Qwen3.5 27B: 39.4 (#59)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Vision | — | 1241 |
Multilingual Inkling leads
Inkling: 54.0 (#52), Qwen3.5 27B: 50.8 (#115)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Non-English | 1434 | 1390 |
| LMArena Chinese | 1490 | 1478 |
| LMArena French | 1458 | 1410 |
| LMArena German | 1446 | 1393 |
| LMArena Japanese | 1429 | 1345 |
| LMArena Korean | 1404 | 1358 |
| LMArena Russian | 1429 | 1390 |
| LMArena Spanish | 1448 | 1407 |
Instruction Following Inkling leads
Inkling: 75.1 (#71), Qwen3.5 27B: 73.5 (#119)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Instruction Following | 1426 | 1393 |
Long Context Too close to call
Inkling: 43.8 (#86), Qwen3.5 27B: 43.1 (#106)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Longer Query | 1434 | 1413 |
Writing & Preference Inkling leads
Inkling: 65.2 (#51), Qwen3.5 27B: 59.3 (#111)
| Benchmark | Inkling | Qwen3.5 27B |
|---|---|---|
| LMArena Text | 1441 | 1409 |
| LMArena Creative Writing | 1387 | 1362 |
| LMArena Multi-Turn | 1436 | 1410 |
| EQ-Bench Creative Writing | 1611 | — |
| EQ-Bench 4 | 1226 | — |
Frequently asked questions
Is Inkling better than Qwen3.5 27B?
Inkling is the stronger model overall, scoring 44.1 to 41.9 on the Noometry Index. Qwen3.5 27B costs 3.1× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Inkling or Qwen3.5 27B?
Qwen3.5 27B is cheaper. It lists at $0.30 per million input tokens and $2.40 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or Qwen3.5 27B better for coding?
Qwen3.5 27B scores higher on coding benchmarks: 38.9 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5 27B does, with 262K tokens against 66K.
How many benchmarks do Inkling and Qwen3.5 27B share?
22 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Qwen3.5 27B has 28.