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 Opus 4.6DeepSeek V3
Claude Opus 4.6

Anthropic

DeepSeek V3

DeepSeek

Pricing
Input $/1M$5.00$0.32
Output $/1M$25.00$0.89
Blended $/1M (3:1)$10.00$0.462
Cache read $/1M$0.50
Cache write $/1M$6.25
Quality indices
Intelligence Index26.48.5
Coding Index23
Math Index26
Agentic Index
Speed & latency
Output speed (tok/s)00
Time to first token0s0s
Time to first answer token0s0s
Benchmarks (Artificial Analysis)
MMLU-Pro75.2%
GPQA Diamond84.0%55.7%
Humanity's Last Exam19.1%2.9%
LiveCodeBench35.9%
SciCode35.8%
MATH-50088.7%
AIME25.3%
AIME 202526.0%
IFBench44.6%34.8%
AA-LCR (Long Context Reasoning)67.0%29.3%
Terminal-Bench Hard48.5%6.8%
Terminal-Bench 2.116.9%
τ²-Bench (Telecom)84.8%22.8%
τ³-Bench Banking4.7%
Design Arena (Elo)
Website1,3031,132
Web apps1,219
Full-stack1,229
Mobile apps1,238
Android native1,216
UI components1,3001,109
Data viz1,2961,103
SVG1,2511,000
3D1,3031,115
Game dev1,3011,077
Agentic game dev1,215
Code categories1,3041,123
ASCII art1,273
Specs
Context window1M164K
Max output tokens128K16K
Input modalitiestext, image, filetext
TokenizerClaudeDeepSeek