Navigate
AI BENCHY
Advertise here

Claude Opus 5 vs OpenAI: GPT-5.6 Luna

The average score is effectively tied at 7.5 vs 7.6. GPT-5.6 Luna (medium) has the lower benchmark cost at $0.352 vs $1.550. GPT-5.6 Luna (medium) is faster at 7.28s vs 7.60s, with pass rates of 57.6% vs 65.2%.

Recommended modelGPT-5.6 Luna (medium)It has the best score here (7.6), while costing about 4.4x less than Claude Opus 5.

Last updated at: 2026-07-25

Metric Claude Opus 5 Claude Opus 5 none Release: 2026-07-25 GPT-5.6 Luna GPT-5.6 Luna medium Release: 2026-07-09
Score 7.5 7.6
Rank #59 #57
Reliability 10.0 10.0
Consistency 9.6 9.6
Tests Correct
Attempt pass rate 57.6% 65.2%
Flaky tests 1 1
Total Runs 66 66
Cost per result 12.912 2.513
Total Cost $1.550 $0.352
Input Price $5.000 / 1M $1.000 / 1M
Output Price $25.000 / 1M $6.000 / 1M
Total Input Tokens 185,550 89,676
Output Tokens 24,866 5,699
Reasoning Tokens 0 37,980
Response Time (avg) 7.60s 7.28s
Response Time (max) 47.72s 29.85s
Response Time (total) 167.28s 160.27s

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#59 Claude Opus 5

none
Cost
$0.271
Time
149.3s
Tokens
10,962 tok

#57 GPT-5.6 Luna

medium
Cost
$0.014
Time
14.6s
Tokens
2,362 tok

Top Models by Score

Score vs Total Cost

Response Time (avg)

Score vs Response Time (avg)

Total Output Tokens

Score vs Total Output Tokens

Category Breakdown

Coding Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
Claude Opus 5 5.9 9.7 33.3% 0 5.54s 10,590 2,886 0
GPT-5.6 Luna 5.4 7.2 44.4% 1 10.38s 7,302 564 9,928

Quick Compare

Switch Comparison Pair