Gemini 3 Flash vs Step 3.7 Flash

Change models
IndexAll categories, out of 100.
54.8 ±3.9 best
54.6 ±5.0
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
95% best
85%
SpeedOutput tokens per second.
198/s best
98/s
EloLMArena rating and rank.
1466
—
ReleasedMonth of release.
2025-12
2026-05
Input / 1MUS dollars per 1M input tokens.
$0.50
$0.20 best
Output / 1MUS dollars per 1M output tokens.
$3.00
$1.15 best
1K chats2K tokens in, 500 out, each.
$2.50
$0.975 best
1K documents20K tokens in, 1K out, each.
$13.00
$5.15 best
1K agent tasks200K tokens in, 20K out, each.
$160.00
$63.00 best
ContextMaximum tokens in one request.
1M best
262K
AgenticMulti-step tasks with tools.
48.3
58.7 best
CodingCode writing and repair.
56.3 best
54.2
ReasoningLogic problems and puzzles.
50.0 best
49.0
MultimodalTasks with images and text.
66.9 best
58.4
KnowledgeFacts and expert knowledge.
60.0 best
46.9
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
40.4
51.3 best
MathMath problems.
56.5
—

Benchmarks

10 shared
Only Gemini 3 Flash: 22Only Step 3.7 Flash: 11
Gemini 3 Flash
Step 3.7 Flash
15.0%123 / 216
21.4% best106 / 216
55.3%141 / 211
73.7% best95 / 211
1.4%108 / 207
2.3% best101 / 207
81.2% best91 / 193
80.9%93 / 193
45.8% best38 / 170
25.8%102 / 170
43.3%98 / 143
98.5% best3 / 143
55.1%68 / 140
67.3% best51 / 140
78.6% best33 / 109
75.3%50 / 109
56.6% best16 / 51
51.6%19 / 51
Claw-EvalAgentic
49.2%38 / 42
67.1% best15 / 42