ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

GLM-5.3-FlashvsStep 3.7 Flash

GLM-5.3-Flash leads 2 of 2 shared benchmarks. GLM-5.3-Flash has the larger context window (1M tokens).

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

Shared benchmarks
2 – 0
GLM-5.3-Flash leads
Cheaper per token
list price, input + output
Larger context
GLM-5.3-Flash
1M tokens
Provider uptime (30d)
— · —
Z.ai · StepFun
GLM-5.3-Flash
Z.ai · released 2026-09-16
textvision
Context
1M
Max out
131K
Input /1M
$0.15
Output /1M
$0.5
Cached /1M
$0.03
Scores
19 · 6 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

BenchmarkGLM-5.3-FlashStep 3.7 FlashΔEdge
Reported by both · 2
HLE (with tools)55.3% 47.2% +8.1 ptGLM-5.3-Flash
Terminal-Bench 2.184.3% 59.6% +24.7 ptGLM-5.3-Flash
Only GLM-5.3-Flash reports · 17
GPQA Diamond90.2% epochnot reported
LMArena Elo1471.9 not reported
Agents' Last Exam (ALE-CLI)26.3% not reported
AutomationBench v1.0.648.8% not reported
BabyVision53.4% not reported
Chartography (with tools)78% not reported
CharXiv Reasoning (with tools)89.4% not reported
DeepSWE v1.163.4% not reported
FrontierMath Tier 4 v2 (Epoch AI run)17.1% epochnot reported
FrontierMath Tiers 1-3 v2 (Epoch AI run)55.8% epochnot reported
GDPval-AA v21773 not reported
MMVU80.5% not reported
MVBench77.8% not reported
NL2Repo56.3% not reported
OfficeQA Pro62.4% not reported
OTIS Mock AIME 2024-2025 (Epoch AI run)93.9% epochnot reported
Toolathlon Verified78.4% not reported
Only Step 3.7 Flash reports · 17
SWE-bench Verifiednot reported76.5%
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%
ResearchRubricsnot reported71.7%
SimpleVQA (with tools)not reported79.2%
SWE-bench Multilingualnot reported72.4%
SWE-bench Pronot reported56.3%
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 GLM-5.3-Flash 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

SpecGLM-5.3-FlashStep 3.7 FlashEdge
Context window1M tokens256K tokensGLM-5.3-Flash
Max output131K tokens
Input price / 1M$0.15
Output price / 1M$0.5
Cached input / 1M$0.03
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$0.65
Modalitiestext · visiontext · visionTie
Released2026-09-162026-05-29
Cited benchmark scores1919
Reliability

Provider status

All providers
More matchups

GLM-5.3-Flash vs …

More matchups

Step 3.7 Flash vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run GLM-5.3-Flash and Step 3.7 Flash on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.