Skip to main content

StepFun: Step 3.7 Flash

StepFun: Step 3.7 Flash is stepfun's model. On the un-leaked meo benchmark it ranks #20 of 22 with 26.6% accuracy and an Effective Value (๐•) of 0.0000 โ€” earning a correct answer for about $0.0256 in 191s each.

as of 2026-06-08

meo rank
#20
of 22
Accuracy
26.6%
overall
๐• index (top=100)
<0.001
raw ๐• 0.0000 ยท N=10
$ / correct
$0.0256
191s each

Per-domain accuracy

Theory of mind
100%
Regex automaton
63%
Perceptual illusions
55%
CSP (unsatisfiable)
50%
Framework bias
33%
Watson-Glaser
25%
Long arithmetic
13%
Logic / math / CS
3%
Base-rate bias
2%
Tape machine
0%
State tracking
0%

Efficiency

Cost per correct
$0.0256
Seconds per correct
191
Tokens per correct
22,065
Correct answers
82 / 251
Total benchmark cost
$2.10
Avg tokens / item
7,209

Aggregated third-party benchmarks

Artificial Analysis (artificialanalysis.ai). Redistribution requires an AA commercial license.

Model details

Lab
stepfun
Released
2026-05-28
Weights
โ€”
Context
256K
Max output
256K
Knowledge cutoff
โ€”
Price in
$0.20/M
Price out
$1.15/M
Modality
text+image+video->text

Sources: openrouter ยท artificial_analysis ยท meo

FAQ

How accurate is StepFun: Step 3.7 Flash?

StepFun: Step 3.7 Flash scores 26.6% overall on the meo benchmark (rank #20 of 22), an un-leaked private holdout graded objective-first.

What is StepFun: Step 3.7 Flash's Effective Value?

Its Effective Value (๐•) is 0.0000 at N=10, ฯ‰=1, ฮด=1.5 โ€” a fused metric of accuracy, speed, cost, and error-cascade. ๐• rankings shift with task depth.

How much does StepFun: Step 3.7 Flash cost per correct answer?

About $0.0256 per correct answer (22,065 tokens, 191 seconds each) at benchmark-time pricing via OpenRouter.

Is StepFun: Step 3.7 Flash open weights?

Open-weights status for StepFun: Step 3.7 Flash is not recorded.