ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

Claude Sonnet 4.6vsGLM-5.3

GLM-5.3 leads 3 of 3 shared benchmarks. GLM-5.3 is 3.1x cheaper per token. Both offer a 1M-token context window.

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

Shared benchmarks
0 – 3
GLM-5.3 leads
Cheaper per token
GLM-5.3
3.1x cheaper, input + output
Larger context
Tie
both 1M tokens
Provider uptime (30d)
100% · —
Anthropic · Z.ai
Claude Sonnet 4.6
Anthropic · released 2026-02-17
textvision
Context
1M
Max out
128K
Input /1M
$3
Output /1M
$15
Cached /1M
$0.3
Scores
19 · 8 core
GLM-5.3
Z.ai · released 2026-08-14
text
Context
1M
Max out
131K
Input /1M
$1.4
Output /1M
$4.4
Cached /1M
$0.26
Scores
33 · 6 core
Quality

Benchmark matrix

BenchmarkClaude Sonnet 4.6GLM-5.3ΔEdge
Reported by both · 3
GPQA Diamond89.9% 90.9% epoch-1 ptGLM-5.3
OTIS Mock AIME 2024-2025 (Epoch AI run)85.8% epoch91.1% epoch-5.3 ptGLM-5.3
SimpleQA Verified35.5% epoch41% epoch-5.5 ptGLM-5.3
Only Claude Sonnet 4.6 reports · 15
SWE-bench Verified79.6% not reported
Humanity's Last Exam (no tools)33.2% not reported
LMArena Elo1458.3 not reported
ARC-AGI-2 (Verified)58.3% not reported
GDPval-AA1633 not reported
Humanity's Last Exam (with tools)49% not reported
MCP Atlas61.3% not reported
MMMLU89.3% not reported
MMMU-Pro (no tools)74.5% not reported
MMMU-Pro (with tools)75.6% not reported
OSWorld-Verified72.5% not reported
SWE-bench Verified (Epoch AI run)75.2% epochnot reported
Tau2-bench Retail91.7% not reported
Tau2-bench Telecom97.9% not reported
Terminal-Bench 2.059.1% not reported
Only GLM-5.3 reports · 21
Agents' Last Examnot reported28.5%
Agents' Last Exam (ALE-CLI)not reported28.5%
AutomationBench v1.0.6not reported48.2%
CyberGymnot reported84.5%
DeepSWE v1.1not reported66.9%
ExploitBenchnot reported54.4%
ExploitGym 2hnot reported105
ExploitGym 6hnot reported130
FrontierMath Tier 4 v2 (Epoch AI run)not reported29.3% epoch
FrontierMath Tiers 1-3 v2 (Epoch AI run)not reported68.8% epoch
FrontierSWEnot reported78.1%
GDPval-AA v2not reported1769
HLE (with tools)not reported62.5%
NL2Reponot reported58%
PostTrainBenchnot reported39.8%
ProgramBench (Almost Solved)not reported19%
SWE-Marathon v1.1not reported42.5%
Terminal-Bench 2.1not reported88.2%
Terminal-Bench 3.0not reported28.3%
Toolathlon Verifiednot reported73%
Z.ai Code Bench (max effort)not reported34.5%
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 Claude Sonnet 4.6 minus GLM-5.3 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

SpecClaude Sonnet 4.6GLM-5.3Edge
Context window1M tokens1M tokensTie
Max output128K tokens131K tokensGLM-5.3
Input price / 1M$3 $1.4 GLM-5.3
Output price / 1M$15 $4.4 GLM-5.3
Cached input / 1M$0.3 $0.26 GLM-5.3
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$18$5.8GLM-5.3
Modalitiestext · visiontextClaude Sonnet 4.6
Released2026-02-172026-08-14
Cited benchmark scores1933
Reliability

Provider status

All providers
More matchups

Claude Sonnet 4.6 vs …

More matchups

GLM-5.3 vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Claude Sonnet 4.6 and GLM-5.3 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.