Model comparison
DeepSeek-V3.2-Exp vs Inkling-Small
Inkling-Small is the stronger model overall, scoring 46.5 to 44.3 on the Noometry Index. DeepSeek-V3.2-Exp costs 2.2× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
DeepSeekSummary
- They share 28 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 6 categories and Inkling-Small in 2 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling-Small leads 38.6 to 22.1.
- The biggest single-benchmark swing is ARC-AGI-2: 4% for DeepSeek-V3.2-Exp and 40.1% for Inkling-Small.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $0.45 / $1.20 for Inkling-Small.
- Inkling-Small accepts more context: 524K tokens versus 164K.
Side by side
| DeepSeek-V3.2-Exp | Inkling-Small | |
|---|---|---|
| Provider | DeepSeek | Thinking Machines Lab |
| Noometry Index | 44.3 | 46.5 |
| Released | 2025-09-29 | 2026-07-15 |
| Weights | Open | Open |
| Context window | 164K | 524K |
| Max output | 66K | 1.05M |
| Input $ / M tokens | $0.26 | $0.45 |
| Output $ / M tokens | $0.38 | $1.20 |
| Results tracked | 49 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), Inkling-Small: 43.6 (#85)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena WebDev | 1362 | 1409 |
| SciCode | 38.9% | 48.7% |
| LMArena Coding | 1454 | 1451 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| SWE-bench Multilingual | 59% | — |
| WeirdML | 39.5% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), Inkling-Small: —
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Inkling-Small leads
DeepSeek-V3.2-Exp: 22.1 (#208), Inkling-Small: 38.6 (#63)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| ARC-AGI-2 | 4% | 40.1% |
| ARC-AGI-1 | 57% | 84% |
| CritPt | 2.9% | 8.3% |
| Chess Puzzles | 14% | 18% |
| LMArena Hard Prompts | 1434 | 1423 |
| Epoch Capabilities Index | 146.27 | 150.15 |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| Thematic Generalization | 65% | — |
| Mystery Game Puzzles | — | 6% |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
Math Inkling-Small leads
DeepSeek-V3.2-Exp: 41.7 (#87), Inkling-Small: 45.1 (#77)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 90% |
| ProofBench | 8% | 6% |
| LMArena Math | 1435 | 1459 |
| FrontierMath (Tiers 1-3) | — | 46.3% |
| FrontierMath Tier 4 | — | 17.1% |
| MathArena Final-Answer Competitions | 57.7% | — |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), Inkling-Small: 48.2 (#77)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| GPQA Diamond | 83.4% | 88.5% |
| LMArena Expert | 1436 | 1442 |
| SimpleQA Verified | — | 19.1% |
| Vectara Hallucination Rate | 5.3% | — |
Multimodal Not comparable
DeepSeek-V3.2-Exp: —, Inkling-Small: 39.1 (#62)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena Vision | — | 1235 |
Multilingual Too close to call
DeepSeek-V3.2-Exp: 52.2 (#90), Inkling-Small: 51.7 (#104)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena Non-English | 1409 | 1402 |
| LMArena Chinese | 1461 | 1465 |
| LMArena French | 1433 | 1436 |
| LMArena German | 1440 | 1405 |
| LMArena Japanese | 1374 | 1405 |
| LMArena Korean | 1371 | 1363 |
| LMArena Russian | 1424 | 1391 |
| LMArena Spanish | 1440 | 1428 |
Instruction Following Too close to call
DeepSeek-V3.2-Exp: 74.5 (#93), Inkling-Small: 73.8 (#114)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena Instruction Following | 1413 | 1399 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), Inkling-Small: 42.7 (#118)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena Longer Query | 1428 | 1401 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Inkling-Small: 59.6 (#107)
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| LMArena Text | 1425 | 1414 |
| LMArena Creative Writing | 1403 | 1331 |
| EQ-Bench Creative Writing | 1515 | 1491 |
| LMArena Multi-Turn | 1427 | 1418 |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Inkling-Small?
Inkling-Small is the stronger model overall, scoring 46.5 to 44.3 on the Noometry Index. DeepSeek-V3.2-Exp costs 2.2× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.2-Exp or Inkling-Small?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Inkling-Small lists at $0.45 and $1.20.
Is DeepSeek-V3.2-Exp or Inkling-Small better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 43.6 in the Noometry coding category.
Which has the bigger context window?
Inkling-Small does, with 524K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and Inkling-Small share?
28 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Inkling-Small has 33.