Model comparison
Gemma 4 26B A4B IT vs Sonar
Gemma 4 26B A4B IT is the stronger model overall, scoring 43.5 to 38.5 on the Noometry Index.
Last verified . 0 shared benchmarks.
Summary
- The widest gap is in math, where Gemma 4 26B A4B IT leads 47.6 to 33.7.
- Gemma 4 26B A4B IT is cheaper at $0.0675 / $0.23 per million input/output tokens, against $1 / $1 for Sonar.
- Gemma 4 26B A4B IT accepts more context: 262K tokens versus 128K.
- Gemma 4 26B A4B IT has downloadable open weights; the other is API-only.
Side by side
| Gemma 4 26B A4B IT | Sonar | |
|---|---|---|
| Provider | Perplexity | |
| Noometry Index | 43.5 | 38.5 |
| Released | 2026-04-02 | 2024-01-01 |
| Weights | Open | Proprietary |
| Context window | 262K | 128K |
| Max output | 33K | 4K |
| Input $ / M tokens | $0.0675 | $1 |
| Output $ / M tokens | $0.23 | $1 |
| Results tracked | 28 | 7 |
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: 39.0 (#164), Sonar: 35.7 (#221)
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LMArena WebDev | 1359 | — |
| SciCode | 40% | — |
| WeirdML | 35.2% | — |
| LiveBench Coding | — | 35.1% |
| LMArena Coding | 1447 | — |
| ALE-Bench | 927.17 | — |
Reasoning Too close to call
Gemma 4 26B A4B IT: 21.8 (#213), Sonar: 21.1 (#227)
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| CritPt | 0% | — |
| Chess Puzzles | 6% | — |
| LiveBench Reasoning | — | 46.3% |
| LMArena Hard Prompts | 1439 | — |
| DTBench | 74.9% | — |
| LiveBench Data Analysis | — | 37.9% |
| LMCA | 29.7% | — |
| Epoch Capabilities Index | 141.85 | — |
| LiveBench | — | 46.9% |
Math Gemma 4 26B A4B IT leads
Gemma 4 26B A4B IT: 47.6 (#67), Sonar: 33.7 (#200)
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 82.2% | — |
| LiveBench Math | — | 41.6% |
| LMArena Math | 1470 | — |
Knowledge Not comparable
Gemma 4 26B A4B IT: 45.7 (#85), Sonar: —
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| GPQA Diamond | 73.2% | — |
| Vectara Hallucination Rate | 5.2% | — |
| LMArena Expert | 1447 | — |
Multimodal Not comparable
Gemma 4 26B A4B IT: 40.6 (#46), Sonar: —
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LMArena Vision | 1260 | — |
Multilingual Not comparable
Gemma 4 26B A4B IT: 53.1 (#71), Sonar: —
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LMArena Non-English | 1421 | — |
| LMArena Chinese | 1495 | — |
| LMArena French | 1460 | — |
| LMArena Russian | 1434 | — |
| LMArena Spanish | 1417 | — |
Instruction Following Gemma 4 26B A4B IT leads
Gemma 4 26B A4B IT: 74.8 (#82), Sonar: 71.4 (#150)
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LiveBench Instruction Following | — | 76.2% |
| LMArena Instruction Following | 1420 | — |
Long Context Not comparable
Gemma 4 26B A4B IT: 43.6 (#91), Sonar: —
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LMArena Longer Query | 1428 | — |
Writing & Preference Gemma 4 26B A4B IT leads
Gemma 4 26B A4B IT: 58.6 (#115), Sonar: 52.6 (#167)
| Benchmark | Gemma 4 26B A4B IT | Sonar |
|---|---|---|
| LMArena Text | 1434 | — |
| LMArena Creative Writing | 1402 | — |
| EQ-Bench Creative Writing | 1305 | — |
| LMArena Multi-Turn | 1441 | — |
| LiveBench Language | — | 44.1% |
Frequently asked questions
Is Gemma 4 26B A4B IT better than Sonar?
Gemma 4 26B A4B IT is the stronger model overall, scoring 43.5 to 38.5 on the Noometry Index.
Which is cheaper, Gemma 4 26B A4B IT or Sonar?
Gemma 4 26B A4B IT is cheaper. It lists at $0.0675 per million input tokens and $0.23 per million output tokens; Sonar lists at $1 and $1.
Is Gemma 4 26B A4B IT or Sonar better for coding?
Gemma 4 26B A4B IT scores higher on coding benchmarks: 39.0 versus 35.7 in the Noometry coding category.
Which has the bigger context window?
Gemma 4 26B A4B IT does, with 262K tokens against 128K.
How many benchmarks do Gemma 4 26B A4B IT and Sonar share?
0 benchmarks have published results for both models. Gemma 4 26B A4B IT has 28 scored results on Noometry and Sonar has 7.