Gemma 4 26B A4B vs GPT-OSS 120B

Change models
IndexAll categories, out of 100.
43.8 ±3.7 best
43.5 ±4.2
GapIndex points below the top.
Top
−0.3 · no clear gap
CoverageIndex weight that has results.
100% best
90%
SpeedOutput tokens per second.
—
224/s
EloLMArena rating and rank.
1434 best
1365
ReleasedMonth of release.
2026-04
2025-08
Input / 1MUS dollars per 1M input tokens.
Free best
$0.09
Output / 1MUS dollars per 1M output tokens.
Free best
$0.36
1K chats2K tokens in, 500 out, each.
Free best
$0.36
1K documents20K tokens in, 1K out, each.
Free best
$2.16
1K agent tasks200K tokens in, 20K out, each.
Free best
$25.20
ContextMaximum tokens in one request.
256K best
131K
AgenticMulti-step tasks with tools.
38.0
41.3 best
CodingCode writing and repair.
44.6 best
42.5
ReasoningLogic problems and puzzles.
41.7 best
40.0
MultimodalTasks with images and text.
49.0
—
KnowledgeFacts and expert knowledge.
43.7 best
42.5
MultilingualTasks in many languages.
36.3
50.9 best
InstructionTasks with strict prompt rules.
55.8 best
52.8
MathMath problems.
56.2 best
51.2

Benchmarks

22 shared
Only Gemma 4 26B A4B: 9Only GPT-OSS 120B: 26
Gemma 4 26B A4B
GPT-OSS 120B
EuroEval PolishMultilingual
55.5%42 / 227
57.6% best29 / 227
73.2%111 / 220
75.8% best106 / 220
19.3%110 / 216
19.6% best109 / 216
EuroEval SwedishMultilingual
58.8%40 / 213
61.3% best32 / 213
65.7% best129 / 211
52.0%146 / 211
EuroEval GermanMultilingual
49.1%46 / 208
51.2% best36 / 208
0.0%145 / 207
1.1% best113 / 207
82.2%84 / 203
88.9% best66 / 203
62.0% best26 / 199
59.3%35 / 199
EuroEval FrenchMultilingual
65.2% best27 / 195
58.4%48 / 195
EuroEval SpanishMultilingual
49.4%51 / 194
53.5% best34 / 194
79.2% best97 / 193
78.2%100 / 193
EuroEval DutchMultilingual
56.4%41 / 192
57.8% best33 / 192
EuroEval ItalianMultilingual
61.1% best28 / 191
58.4%35 / 191
18.4%139 / 170
21.7% best123 / 170
82.3% best64 / 144
80.1%67 / 144
43.6%97 / 143
65.8% best76 / 143
72.4% best35 / 140
69.0%45 / 140
3.2%97 / 132
5.4% best95 / 132
19.6%117 / 120
21.1% best109 / 120