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 4.6MiniMax M3 (batch)
Claude Opus 4.6

Anthropic

MiniMax M3 (batch)

MiniMax

Pricing
Input $/1M$5.00$0.30
Output $/1M$25.00$1.20
Blended $/1M (3:1)$10.00$0.525
Cache read $/1M$0.50$0.06
Cache write $/1M$6.25
Quality indices
Intelligence Index26.429.2
Coding Index58.6
Math Index
Agentic Index29.5
Speed & latency
Output speed (tok/s)0107
Time to first token0s1.01s
Time to first answer token0s19.78s
Benchmarks (Artificial Analysis)
GPQA Diamond84.0%92.9%
Humanity's Last Exam19.1%39.0%
SciCode47.1%
IFBench44.6%82.9%
AA-LCR (Long Context Reasoning)67.0%83.0%
Terminal-Bench Hard48.5%42.4%
Terminal-Bench 2.165.2%
τ²-Bench (Telecom)84.8%88.9%
τ³-Bench Banking15.3%
Design Arena (Elo)
Website1,3031,272
Web apps1,2191,215
Full-stack1,2291,204
Mobile apps1,2381,195
Android native1,2161,167
UI components1,3001,261
Data viz1,2961,250
SVG1,2511,191
3D1,3031,238
Game dev1,3011,239
Agentic game dev1,2151,158
HTML slides1,183
Python→PPTX slides1,207
Code categories1,3041,265
ASCII art1,2731,183
Specs
Context window1M524K
Max output tokens128K472K
Input modalitiestext, image, filetext, image, video
TokenizerClaudeOther