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 5Step 3.7 Flash
Claude Sonnet 5

Anthropic

Step 3.7 Flash

Stepfun

Pricing
Input $/1M$2.00$0.20
Output $/1M$10.00$1.15
Blended $/1M (3:1)$4.00$0.438
Cache read $/1M$0.20$0.04
Cache write $/1M$2.50
Quality indices
Intelligence Index38.219.5
Coding Index71.539.6
Math Index
Agentic Index43.6
Speed & latency
Output speed (tok/s)82179
Time to first token118.14s1.39s
Time to first answer token118.14s12.59s
Benchmarks (Artificial Analysis)
GPQA Diamond91.1%80.9%
Humanity's Last Exam41.3%21.4%
SciCode54.3%43.9%
IFBench67.3%
AA-LCR (Long Context Reasoning)82.0%73.7%
Terminal-Bench Hard35.6%
Terminal-Bench 2.180.5%39.3%
τ²-Bench (Telecom)98.5%
τ³-Bench Banking37.3%12.0%
Design Arena (Elo)
Website1,2881,207
Web apps1,260
Full-stack1,260
Mobile apps1,249
Android native1,240
UI components1,2991,198
Data viz1,2601,185
SVG1,2121,099
3D1,2891,157
Game dev1,3131,181
Agentic game dev1,227
Godot game dev1,238
HTML slides1,220
Python→PPTX slides1,216
Code categories1,2931,196
ASCII art1,2261,167
Specs
Context window1M262K
Max output tokens128K230K
Input modalitiestext, image, filetext, image, video
TokenizerClaudeOther