ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

LongCat Flash ChatvsStep 3.7 Flash

Step 3.7 Flash leads 1 of 1 shared benchmark. Step 3.7 Flash has the larger context window (256K tokens).

Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.

Shared benchmarks
0 – 1
Step 3.7 Flash leads
Cheaper per token
list price, input + output
Larger context
Step 3.7 Flash
256K tokens
Provider uptime (30d)
— · —
Meituan LongCat · StepFun
LongCat Flash Chat
Meituan LongCat · released 2025-09-01
text
Context
128K
Max out
Input /1M
Output /1M
Cached /1M
Scores
31 · 9 core
Step 3.7 Flash
StepFun · released 2026-05-29
textvision
Context
256K
Max out
Input /1M
Output /1M
Cached /1M
Scores
19 · 6 core
Quality

Benchmark matrix

BenchmarkLongCat Flash ChatStep 3.7 FlashΔEdge
Reported by both · 1
SWE-bench Verified60.4% 76.5% -16.1 ptStep 3.7 Flash
Only LongCat Flash Chat reports · 30
GPQA Diamond73.23% not reported
MMLU89.71% not reported
MMLU-Pro82.68% not reported
AIME 202561.25% not reported
LMArena Elo1422.9 not reported
AceBench76.1% not reported
AIME 202470.42% not reported
ArenaHard-V286.5% not reported
BeyondAIME43% not reported
C-Eval90.44% not reported
CMMLU84.34% not reported
COLLIE57.1% not reported
DROP79.06% not reported
GraphWalks-128k51.05% not reported
HumanEval+88.41% not reported
IFEval89.65% not reported
LiveCodeBench48.02% not reported
MATH50096.4% not reported
MBPP+79.63% not reported
Meeseeks-zh43.03% not reported
Safety: Criminal91.24% not reported
Safety: Harmful83.98% not reported
Safety: Misinformation81.72% not reported
Safety: Privacy93.98% not reported
Tau2-Bench (airline)58% not reported
Tau2-Bench (retail)71.27% not reported
Tau2-Bench (telecom)73.68% not reported
TerminalBench39.51% not reported
VitaBench24.3% not reported
ZebraLogic89.3% not reported
Only Step 3.7 Flash reports · 18
AIME 2026not reported95% matharena
AA-LCRnot reported63.9%
BrowseCompnot reported75.8%
ClawEval v1.1not reported67.1%
DeepSearchQA Accuracynot reported81.7%
DeepSearchQA F1not reported92.8%
GDPval-Stirrupnot reported1415.8
GDPval-Stirrup Intelligence Indexnot reported45.8
HLE (with tools, text-only)not reported49.7%
HLE (with tools)not reported47.2%
ResearchRubricsnot reported71.7%
SimpleVQA (with tools)not reported79.2%
SWE-bench Multilingualnot reported72.4%
SWE-bench Pronot reported56.3%
Terminal-Bench 2.1not reported59.6%
Toolathlonnot reported49.5%
V* (with Python)not reported95.3%
WorldVQA (with visual search)not reported58.1%
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is LongCat Flash Chat minus Step 3.7 Flash in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing

Side by side

SpecLongCat Flash ChatStep 3.7 FlashEdge
Context window128K tokens256K tokensStep 3.7 Flash
Max output
Input price / 1M
Output price / 1M
Cached input / 1M
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
Modalitiestexttext · visionStep 3.7 Flash
Released2025-09-012026-05-29
Cited benchmark scores3119
Reliability

Provider status

All providers
More matchups

LongCat Flash Chat vs …

More matchups

Step 3.7 Flash vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run LongCat Flash Chat and Step 3.7 Flash on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.