Model comparison
Gemma 3 27B vs Inkling
Inkling is the stronger model overall, scoring 44.1 to 30.8 on the Noometry Index. Gemma 3 27B costs 26× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
GoogleSummary
- They share 26 benchmarks with published results for both. Gemma 3 27B scores higher in 0 categories and Inkling in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Inkling leads 55.1 to 25.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 22.5% for Gemma 3 27B and 88.9% for Inkling.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Gemma 3 27B accepts more context: 131K tokens versus 66K.
Side by side
| Gemma 3 27B | Inkling | |
|---|---|---|
| Provider | Thinking Machines Lab | |
| Noometry Index | 30.8 | 44.1 |
| Released | 2025-03-11 | 2026-07-15 |
| Weights | Open | Open |
| Context window | 131K | 66K |
| Max output | 8K | 66K |
| Input $ / M tokens | $0.08 | $1.87 |
| Output $ / M tokens | $0.16 | $4.68 |
| Results tracked | 43 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling leads
Gemma 3 27B: 22.5 (#334), Inkling: 34.5 (#234)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| SciCode | 21.2% | 47% |
| LMArena Coding | 1322 | 1464 |
| FrontierCode | — | 14% |
| Aider Polyglot | 4.9% | — |
| LMArena WebDev | — | 1413 |
| FrontierSWE | — | 4.1% |
| WeirdML | — | 32.3% |
| LiveBench Coding | 39.9% | — |
| ALE-Bench | — | 946 |
Agentic & Tool Use Inkling leads
Gemma 3 27B: 25.1 (#110), Inkling: 29.6 (#85)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| APEX-Agents | — | 33.8% |
| Berkeley Function Calling Leaderboard | 29.5% | — |
| τ²-bench Banking | — | 25% |
Reasoning Inkling leads
Gemma 3 27B: 16.7 (#301), Inkling: 40.4 (#56)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| CritPt | 0% | 5.4% |
| Chess Puzzles | 0% | 21% |
| LMArena Hard Prompts | 1340 | 1451 |
| DTBench | 52.5% | 87.5% |
| LMCA | 12.3% | 37.6% |
| Epoch Capabilities Index | 130.04 | 148.54 |
| ARC-AGI-2 | — | 36.5% |
| SimpleBench | — | 50% |
| Kagi LLM Benchmark | 40.4% | — |
| ARC-AGI-1 | — | 79.5% |
| LiveBench Reasoning | 43.8% | — |
| LiveBench Data Analysis | 51.5% | — |
| LiveBench | 50% | — |
Math Inkling leads
Gemma 3 27B: 25.9 (#265), Inkling: 31.3 (#225)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 88.9% |
| LMArena Math | 1312 | 1479 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | — | 0% |
| LiveBench Math | 55.4% | — |
| MATH Level 5 | 74% | — |
Knowledge Inkling leads
Gemma 3 27B: 25.5 (#261), Inkling: 55.1 (#49)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| GPQA Diamond | 47.7% | 88.3% |
| LMArena Expert | 1304 | 1465 |
| SimpleQA Verified | — | 40.3% |
| Confabulations | 40.3% | — |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), Inkling: —
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Inkling leads
Gemma 3 27B: 46.9 (#155), Inkling: 54.0 (#52)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| LMArena Non-English | 1334 | 1434 |
| LMArena Chinese | 1346 | 1490 |
| LMArena French | 1368 | 1458 |
| LMArena German | 1362 | 1446 |
| LMArena Japanese | 1287 | 1429 |
| LMArena Korean | 1308 | 1404 |
| LMArena Russian | 1349 | 1429 |
| LMArena Spanish | 1349 | 1448 |
Instruction Following Inkling leads
Gemma 3 27B: 70.6 (#160), Inkling: 75.1 (#71)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| LMArena Instruction Following | 1321 | 1426 |
| LiveBench Instruction Following | 74.9% | — |
Long Context Inkling leads
Gemma 3 27B: 27.6 (#293), Inkling: 43.8 (#86)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| LMArena Longer Query | 1333 | 1434 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Inkling leads
Gemma 3 27B: 52.5 (#168), Inkling: 65.2 (#51)
| Benchmark | Gemma 3 27B | Inkling |
|---|---|---|
| LMArena Text | 1358 | 1441 |
| LMArena Creative Writing | 1346 | 1387 |
| EQ-Bench Creative Writing | 1266 | 1611 |
| LMArena Multi-Turn | 1345 | 1436 |
| Short-Story Creative Writing | 79.9% | — |
| EQ-Bench 4 | — | 1226 |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than Inkling?
Inkling is the stronger model overall, scoring 44.1 to 30.8 on the Noometry Index. Gemma 3 27B costs 26× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or Inkling?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Gemma 3 27B or Inkling better for coding?
Inkling scores higher on coding benchmarks: 34.5 versus 22.5 in the Noometry coding category.
Which has the bigger context window?
Gemma 3 27B does, with 131K tokens against 66K.
How many benchmarks do Gemma 3 27B and Inkling share?
26 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Inkling has 41.