About
Step 3.5 Flash is StepFun's most capable open-source foundation model. Built on a sparse Mixture of Experts (MoE) architecture, it selectively activates only 11B of its 196B parameters per token....
Related Models
Pricing
Input
$0.09
per 1M tokens
Output
$0.30
per 1M tokens
Blended
$0.14
per 1M tokens
Cheaper than 77% of models. Median price is $0.54/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.14
Monthly
$4.28
vs. Similar Models
GPT-5.2 (Non-reasoning)Q:0.0
$4.81+3278%
Hy3-preview (Non-reasoning)Q:+0.1
$0.20+40%
inclusionAI: Ling-2.6-1TQ:+0.1
$0.21+49%
Google: Gemini 2.5 ProQ:-0.2
$3.44+2312%
Performance
194
tokens/sec
Faster than 85% of models
1.19
seconds
Faster than 45% of models
11.52
seconds
Faster than 38% of models
Market Median
95 tok/s
105% faster
Median TTFT
1.11s
8% slower
Throughput/Dollar
1359
tok/s per $/1M
Speed Comparison
inclusionAI: Ling-2.6-flash
194 tok/s+0%
Qwen: Qwen3 Next 80B A3B Instruct
193 tok/s-1%
Step 3.5 Flash
193 tok/s-1%
Context Window
262K
tokens
Larger than 62% of models
Max Output
16K
tokens
6% of context
Benchmarks
MMLU-ProNot evaluated
GPQA Diamond
82.6%
HLE
22.6%
LiveCodeBenchNot evaluated
SciCode
38.5%
TerminalBench Hard
32.6%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
66.5%
Long Context Recall
54.3%
Tau2
87.4%
Market AverageTop Score