ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

Claude Sonnet 4.6vsGPT-5.4

GPT-5.4 leads 12 of 13 shared benchmarks. GPT-5.4 is marginally cheaper per token. GPT-5.4 has the larger context window (1.1M tokens).

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

Shared benchmarks
1 – 12
GPT-5.4 leads
Cheaper per token
GPT-5.4
1x cheaper, input + output
Larger context
GPT-5.4
1.1M tokens
Provider uptime (30d)
100% · 95.5679%
Anthropic · OpenAI
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
GPT-5.4
OpenAI · released 2026-03-05
textvision
Context
1.1M
Max out
128K
Input /1M
$2.5
Output /1M
$15
Cached /1M
$0.25
Scores
21 · 11 core
Quality

Benchmark matrix

BenchmarkClaude Sonnet 4.6GPT-5.4ΔEdge
Reported by both · 13
GPQA Diamond89.9% 92.8% -2.9 ptGPT-5.4
Humanity's Last Exam (no tools)33.2% 39.8% -6.6 ptGPT-5.4
LMArena Elo1458.3 1452.6 +5.7Claude Sonnet 4.6
Humanity's Last Exam (with tools)49% 52.1% -3.1 ptGPT-5.4
MCP Atlas61.3% 67.2% -5.9 ptGPT-5.4
MMMU-Pro (no tools)74.5% 81.2% -6.7 ptGPT-5.4
MMMU-Pro (with tools)75.6% 82.1% -6.5 ptGPT-5.4
OSWorld-Verified72.5% 75% -2.5 ptGPT-5.4
OTIS Mock AIME 2024-2025 (Epoch AI run)85.8% epoch97.8% epoch-12 ptGPT-5.4
SimpleQA Verified35.5% epoch45.1% epoch-9.6 ptGPT-5.4
SWE-bench Verified (Epoch AI run)75.2% epoch76.9% epoch-1.7 ptGPT-5.4
Tau2-bench Telecom97.9% 98.9% -1 ptGPT-5.4
Terminal-Bench 2.059.1% 75.1% -16 ptGPT-5.4
Only Claude Sonnet 4.6 reports · 5
SWE-bench Verified79.6% not reported
ARC-AGI-2 (Verified)58.3% not reported
GDPval-AA1633 not reported
MMMLU89.3% not reported
Tau2-bench Retail91.7% not reported
Only GPT-5.4 reports · 7
AIME 2026not reported99.2% matharena
BrowseCompnot reported82.7%
FrontierMath Tier 4 v2 (Epoch AI run)not reported49% epoch
FrontierMath Tiers 1-3 v2 (Epoch AI run)not reported78.6% epoch
GDPval (wins or ties)not reported83%
SWE-Bench Pronot reported57.7%
Toolathlonnot reported54.6%
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 GPT-5.4 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 4.6GPT-5.4Edge
Context window1M tokens1.1M tokensGPT-5.4
Max output128K tokens128K tokensTie
Input price / 1M$3 $2.5 GPT-5.4
Output price / 1M$15 $15 Tie
Cached input / 1M$0.3 $0.25 GPT-5.4
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$18$17.5GPT-5.4
Modalitiestext · visiontext · visionTie
Released2026-02-172026-03-05
Cited benchmark scores1921
Reliability

Provider status

All providers
More matchups

Claude Sonnet 4.6 vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Claude Sonnet 4.6 and GPT-5.4 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.