AiCostCompare

Compare models

Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.

Claude Sonnet 5DeepSeek V3.2
Claude Sonnet 5

Anthropic

DeepSeek V3.2

DeepSeek

Pricing
Input $/1M$2.00$0.269
Output $/1M$10.00$0.40
Blended $/1M (3:1)$4.00$0.302
Cache read $/1M$0.20$0.134
Cache write $/1M$2.50
Quality indices
Intelligence Index53.424.7
Coding Index71.544.2
Math Index59
Agentic Index46.718.3
Speed & latency
Output speed (tok/s)880
Time to first token93.55s0s
Time to first answer token93.55s0s
Benchmarks (Artificial Analysis)
MMLU-Pro83.7%
GPQA Diamond91.1%75.1%
Humanity's Last Exam39.6%10.5%
LiveCodeBench59.3%
SciCode53.6%38.7%
AIME 202559.0%
IFBench49.0%
AA-LCR (Long Context Reasoning)70.7%39.0%
Terminal-Bench Hard32.6%
Terminal-Bench 2.180.5%
τ²-Bench (Telecom)78.9%
τ³-Bench Banking28.2%
Design Arena (Elo)
Website1,3141,201
Web apps1,305
Full-stack1,292
Android native1,255
UI components1,3171,190
Data viz1,2811,196
SVG1,2391,079
3D1,3211,192
Game dev1,3451,191
Agentic game dev1,241
Godot game dev1,257
HTML slides1,233
Python→PPTX slides1,252
Code categories1,3121,197
ASCII art1,2301,124
Specs
Context window1M164K
Max output tokens128K66K
Input modalitiestext, image, filetext
TokenizerClaudeDeepSeek