Qwen3.5 397B (Reasoning) vs Step 3.7 Flash
Change modelsQwen3.5 397B (Reasoning)Alibaba
Step 3.7 FlashStepFun
IndexAll categories, out of 100.
55.3 ±6.9
55.3 ±5.0
GapIndex points below the top.
Top
Top
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
—
112/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-02
2026-05
Input / 1MUS dollars per 1M input tokens.
$0.60
$0.20 best
Output / 1MUS dollars per 1M output tokens.
$3.60
$1.15 best
1K chats2K tokens in, 500 out, each.
$3.00
$0.975 best
1K documents20K tokens in, 1K out, each.
$15.60
$5.15 best
1K agent tasks200K tokens in, 20K out, each.
$192.00
$63.00 best
ContextMaximum tokens in one request.
128K
262K best
AgenticMulti-step tasks with tools.
59.4
59.6 best
CodingCode writing and repair.
54.2
54.9 best
ReasoningLogic problems and puzzles.
50.2 best
49.6
MultimodalTasks with images and text.
55.8
59.3 best
KnowledgeFacts and expert knowledge.
56.3 best
46.8
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
62.7 best
51.9
MathMath problems.
—
—
Benchmarks
9 sharedOnly Step 3.7 Flash: 12
Qwen3.5 397B (Reasoning)
Step 3.7 Flash