Model comparison
DeepSeek-V3.1 vs Inkling-Small
Inkling-Small is the stronger model overall, scoring 46.5 to 42.2 on the Noometry Index. DeepSeek-V3.1 costs 1.5× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Last verified . 19 shared benchmarks.
DeepSeekSummary
- They share 19 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 3 categories and Inkling-Small in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling-Small leads 39.0 to 27.3.
- DeepSeek-V3.1 is cheaper at $0.25 / $0.95 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-V3.1 | Inkling-Small | |
|---|---|---|
| Provider | DeepSeek | Thinking Machines Lab |
| Noometry Index | 42.2 | 46.5 |
| Released | 2025-08-21 | 2026-07-15 |
| Weights | Open | Open |
| Context window | 164K | 524K |
| Max output | 8K | 1.05M |
| Input $ / M tokens | $0.25 | $0.45 |
| Output $ / M tokens | $0.95 | $1.20 |
| Results tracked | 27 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling-Small leads
DeepSeek-V3.1: 39.8 (#140), Inkling-Small: 43.3 (#80)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Coding | 1417 | 1451 |
| LMArena WebDev | — | 1408 |
| SciCode | — | 48.7% |
| WeirdML | 38.4% | — |
Reasoning Inkling-Small leads
DeepSeek-V3.1: 27.3 (#109), Inkling-Small: 39.0 (#53)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Hard Prompts | 1417 | 1424 |
| Epoch Capabilities Index | 139.92 | 150.13 |
| ARC-AGI-2 | — | 40.1% |
| SimpleBench | 40% | — |
| Kagi LLM Benchmark | 53.2% | — |
| ARC-AGI-1 | — | 84% |
| CritPt | — | 8.3% |
| Chess Puzzles | — | 18% |
| Mystery Game Puzzles | — | 6% |
| DTBench | 82.7% | — |
| LMCA | 24.3% | — |
| ForecastBench | 58 | — |
Math Inkling-Small leads
DeepSeek-V3.1: 36.1 (#125), Inkling-Small: 44.9 (#65)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Math | 1420 | 1453 |
| FrontierMath (Tiers 1-3) | — | 46.3% |
| FrontierMath Tier 4 | — | 17.1% |
| OTIS Mock AIME 2024-2025 | — | 90% |
| ProofBench | — | 6% |
Knowledge Inkling-Small leads
DeepSeek-V3.1: 43.0 (#86), Inkling-Small: 48.3 (#71)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Expert | 1403 | 1442 |
| GPQA Diamond | — | 88.5% |
| SimpleQA Verified | — | 19.1% |
| Vectara Hallucination Rate | 5.5% | — |
Multimodal Not comparable
DeepSeek-V3.1: —, Inkling-Small: 38.4 (#59)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Vision | — | 1235 |
Multilingual Too close to call
DeepSeek-V3.1: 51.8 (#100), Inkling-Small: 51.7 (#102)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Non-English | 1400 | 1400 |
| LMArena Chinese | 1470 | 1458 |
| LMArena French | 1447 | 1438 |
| LMArena German | 1412 | 1412 |
| LMArena Japanese | 1374 | 1399 |
| LMArena Korean | 1337 | 1355 |
| LMArena Russian | 1405 | 1392 |
| LMArena Spanish | 1429 | 1426 |
Instruction Following Too close to call
DeepSeek-V3.1: 74.1 (#105), Inkling-Small: 73.9 (#112)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Instruction Following | 1400 | 1397 |
Long Context Inkling-Small leads
DeepSeek-V3.1: 36.3 (#227), Inkling-Small: 42.7 (#113)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Longer Query | 1422 | 1401 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference Too close to call
DeepSeek-V3.1: 60.5 (#93), Inkling-Small: 59.8 (#102)
| Benchmark | DeepSeek-V3.1 | Inkling-Small |
|---|---|---|
| LMArena Text | 1420 | 1413 |
| LMArena Creative Writing | 1401 | 1330 |
| EQ-Bench Creative Writing | 1436 | 1491 |
| LMArena Multi-Turn | 1408 | 1420 |
Frequently asked questions
Is DeepSeek-V3.1 better than Inkling-Small?
Inkling-Small is the stronger model overall, scoring 46.5 to 42.2 on the Noometry Index. DeepSeek-V3.1 costs 1.5× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.1 or Inkling-Small?
DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 per million output tokens; Inkling-Small lists at $0.45 and $1.20.
Is DeepSeek-V3.1 or Inkling-Small better for coding?
Inkling-Small scores higher on coding benchmarks: 43.3 versus 39.8 in the Noometry coding category.
Which has the bigger context window?
Inkling-Small does, with 524K tokens against 164K.
How many benchmarks do DeepSeek-V3.1 and Inkling-Small share?
19 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Inkling-Small has 33.