ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

Claude Sonnet 4.6vsDeepSeek V4 Pro

DeepSeek V4 Pro leads 5 of 6 shared benchmarks. DeepSeek V4 Pro is 13.8x 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
1 – 5
DeepSeek V4 Pro leads
Cheaper per token
DeepSeek V4 Pro
13.8x cheaper, input + output
Larger context
Tie
both 1M tokens
Provider uptime (30d)
100% · 100%
Anthropic · DeepSeek
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
DeepSeek V4 Pro
DeepSeek · released 2026-08-13
text
Context
1M
Max out
384K
Input /1M
$0.435
Output /1M
$0.87
Cached /1M
$0.004
Scores
11 · 9 core
Quality

Benchmark matrix

BenchmarkClaude Sonnet 4.6DeepSeek V4 ProΔEdge
Reported by both · 6
GPQA Diamond89.9% 90.9% epoch-1 ptDeepSeek V4 Pro
LMArena Elo1458.3 1450.6 +7.7Claude Sonnet 4.6
Humanity's Last Exam (with tools)49% 60% -11 ptDeepSeek V4 Pro
OTIS Mock AIME 2024-2025 (Epoch AI run)85.8% epoch96.7% epoch-10.9 ptDeepSeek V4 Pro
SimpleQA Verified35.5% epoch47% epoch-11.5 ptDeepSeek V4 Pro
SWE-bench Verified (Epoch AI run)75.2% epoch77.6% epoch-2.4 ptDeepSeek V4 Pro
Only Claude Sonnet 4.6 reports · 12
SWE-bench Verified79.6% not reported
Humanity's Last Exam (no tools)33.2% not reported
ARC-AGI-2 (Verified)58.3% not reported
GDPval-AA1633 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
Tau2-bench Retail91.7% not reported
Tau2-bench Telecom97.9% not reported
Terminal-Bench 2.059.1% not reported
Only DeepSeek V4 Pro reports · 5
AIME 2026not reported96.7% matharena
DeepSWE v1.1not reported62.7%
FrontierMath Tier 4 v2 (Epoch AI run)not reported2.4% epoch
FrontierMath Tiers 1-3 v2 (Epoch AI run)not reported45.3% epoch
Terminal-Bench 2.1not reported87.9%
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 DeepSeek V4 Pro 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 4.6DeepSeek V4 ProEdge
Context window1M tokens1M tokensTie
Max output128K tokens384K tokensDeepSeek V4 Pro
Input price / 1M$3 $0.435 DeepSeek V4 Pro
Output price / 1M$15 $0.87 DeepSeek V4 Pro
Cached input / 1M$0.3 $0.004 DeepSeek V4 Pro
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$18$1.31DeepSeek V4 Pro
Modalitiestext · visiontextClaude Sonnet 4.6
Released2026-02-172026-08-13
Cited benchmark scores1911
Reliability

Provider status

All providers
More matchups

Claude Sonnet 4.6 vs …

More matchups

DeepSeek V4 Pro vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Claude Sonnet 4.6 and DeepSeek V4 Pro on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.