Model comparison
Gemma 4 26B A4B IT vs Inkling
Inkling is the stronger model overall, scoring 44.1 to 43.1 on the Noometry Index. Gemma 4 26B A4B IT costs 21× 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 4 26B A4B IT scores higher in 2 categories and Inkling in 6 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling leads 40.6 to 21.1.
- The biggest single-benchmark swing is GPQA Diamond: 73.2% for Gemma 4 26B A4B IT and 88.3% for Inkling.
- Gemma 4 26B A4B IT is cheaper at $0.0765 / $0.26 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Gemma 4 26B A4B IT accepts more context: 262K tokens versus 66K.
Side by side
| Gemma 4 26B A4B IT | Inkling | |
|---|---|---|
| Provider | Thinking Machines Lab | |
| Noometry Index | 43.1 | 44.1 |
| Released | 2026-04-02 | 2026-07-15 |
| Weights | Open | Open |
| Context window | 262K | 66K |
| Max output | 33K | 66K |
| Input $ / M tokens | $0.0765 | $1.87 |
| Output $ / M tokens | $0.26 | $4.68 |
| Results tracked | 28 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 4 26B A4B IT leads
Gemma 4 26B A4B IT: 38.8 (#156), Inkling: 34.1 (#229)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena WebDev | 1359 | 1412 |
| SciCode | 40% | 47% |
| WeirdML | 35.2% | 32.3% |
| LMArena Coding | 1446 | 1463 |
| ALE-Bench | 927.17 | 946 |
| FrontierCode | — | 14% |
| FrontierSWE | — | 4.1% |
Agentic & Tool Use Not comparable
Gemma 4 26B A4B IT: —, Inkling: 29.5 (#76)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| APEX-Agents | — | 33.8% |
| τ²-bench Banking | — | 25% |
Reasoning Inkling leads
Gemma 4 26B A4B IT: 21.1 (#205), Inkling: 40.6 (#48)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| CritPt | 0% | 5.4% |
| Chess Puzzles | 6% | 21% |
| LMArena Hard Prompts | 1438 | 1452 |
| DTBench | 74.9% | 87.5% |
| LMCA | 29.7% | 37.6% |
| Epoch Capabilities Index | 141.85 | 148.6 |
| ARC-AGI-2 | — | 36.5% |
| SimpleBench | — | 50% |
| ARC-AGI-1 | — | 79.5% |
Math Gemma 4 26B A4B IT leads
Gemma 4 26B A4B IT: 45.7 (#61), Inkling: 31.2 (#198)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 82.2% | 88.9% |
| LMArena Math | 1469 | 1482 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | — | 0% |
Knowledge Inkling leads
Gemma 4 26B A4B IT: 45.6 (#79), Inkling: 55.2 (#44)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| GPQA Diamond | 73.2% | 88.3% |
| LMArena Expert | 1447 | 1468 |
| SimpleQA Verified | — | 40.3% |
| Vectara Hallucination Rate | 5.2% | — |
Multimodal Not comparable
Gemma 4 26B A4B IT: 39.9 (#43), Inkling: —
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena Vision | 1260 | — |
Multilingual Too close to call
Gemma 4 26B A4B IT: 53.3 (#68), Inkling: 54.2 (#49)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena Non-English | 1422 | 1434 |
| LMArena Chinese | 1495 | 1495 |
| LMArena French | 1459 | 1453 |
| LMArena Russian | 1434 | 1430 |
| LMArena Spanish | 1415 | 1442 |
| LMArena German | — | 1456 |
| LMArena Japanese | — | 1428 |
| LMArena Korean | — | 1401 |
Instruction Following Too close to call
Gemma 4 26B A4B IT: 75.0 (#78), Inkling: 75.3 (#65)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena Instruction Following | 1420 | 1427 |
Long Context Too close to call
Gemma 4 26B A4B IT: 43.6 (#87), Inkling: 43.9 (#79)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena Longer Query | 1427 | 1436 |
Writing & Preference Inkling leads
Gemma 4 26B A4B IT: 58.8 (#110), Inkling: 65.5 (#48)
| Benchmark | Gemma 4 26B A4B IT | Inkling |
|---|---|---|
| LMArena Text | 1434 | 1443 |
| LMArena Creative Writing | 1401 | 1390 |
| EQ-Bench Creative Writing | 1305 | 1611 |
| LMArena Multi-Turn | 1442 | 1436 |
| EQ-Bench 4 | — | 1226 |
Frequently asked questions
Is Gemma 4 26B A4B IT better than Inkling?
Inkling is the stronger model overall, scoring 44.1 to 43.1 on the Noometry Index. Gemma 4 26B A4B IT costs 21× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Gemma 4 26B A4B IT or Inkling?
Gemma 4 26B A4B IT is cheaper. It lists at $0.0765 per million input tokens and $0.26 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Gemma 4 26B A4B IT or Inkling better for coding?
Gemma 4 26B A4B IT scores higher on coding benchmarks: 38.8 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
Gemma 4 26B A4B IT does, with 262K tokens against 66K.
How many benchmarks do Gemma 4 26B A4B IT and Inkling share?
26 benchmarks have published results for both models. Gemma 4 26B A4B IT has 28 scored results on Noometry and Inkling has 41.