Gemma 4 26B A4B vs Qwen3.5-4B
Change modelsGemma 4 26B A4BGoogle
Qwen3.5-4BAlibaba
IndexAll categories, out of 100.
44.5 ±3.9 best
44.3 ±7.4
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
100% best
80%
SpeedOutput tokens per second.
—
20/s
EloLMArena rating and rank.
1434
—
ReleasedMonth of release.
2026-04
—
Input / 1MUS dollars per 1M input tokens.
Free
—
Output / 1MUS dollars per 1M output tokens.
Free
—
1K chats2K tokens in, 500 out, each.
Free
—
1K documents20K tokens in, 1K out, each.
Free
—
1K agent tasks200K tokens in, 20K out, each.
Free
—
ContextMaximum tokens in one request.
256K
262K best
AgenticMulti-step tasks with tools.
34.0
63.9 best
CodingCode writing and repair.
45.3
—
ReasoningLogic problems and puzzles.
42.1
43.7 best
MultimodalTasks with images and text.
49.6 best
40.3
KnowledgeFacts and expert knowledge.
46.7 best
41.1
MultilingualTasks in many languages.
36.3 best
35.5
InstructionTasks with strict prompt rules.
56.7 best
37.6
MathMath problems.
56.8 best
41.4
Benchmarks
16 sharedOnly Gemma 4 26B A4B: 13
Gemma 4 26B A4B
Qwen3.5-4B