Head-to-head comparison across 1benchmark categories. Overall scores shown here use BenchLM's provisional ranking lane.
GPT-5.5 Pro
100
Step 3.7 Flash
72
Pick GPT-5.5 Pro if you want the stronger benchmark profile. Step 3.7 Flash only becomes the better choice if you want the cheaper token bill.
Agentic
+24.2 difference
GPT-5.5 Pro
Step 3.7 Flash
$30 / $180
$0.2 / $1.15
N/A
N/A
N/A
N/A
1M
256K
Pick GPT-5.5 Pro if you want the stronger benchmark profile. Step 3.7 Flash only becomes the better choice if you want the cheaper token bill.
GPT-5.5 Pro is clearly ahead on the provisional aggregate, 100 to 72. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.5 Pro's sharpest advantage is in agentic, where it averages 90.1 against 65.9. The single biggest benchmark swing on the page is BrowseComp, 90.1% to 75.8%.
GPT-5.5 Pro is also the more expensive model on tokens at $30.00 input / $180.00 output per 1M tokens, versus $0.20 input / $1.15 output per 1M tokens for Step 3.7 Flash. That is roughly 156.5x on output cost alone. GPT-5.5 Pro gives you the larger context window at 1M, compared with 256K for Step 3.7 Flash.
GPT-5.5 Pro is ahead on BenchLM's provisional leaderboard, 100 to 72. The biggest single separator in this matchup is BrowseComp, where the scores are 90.1% and 75.8%.
GPT-5.5 Pro has the edge for agentic tasks in this comparison, averaging 90.1 versus 65.9. Inside this category, BrowseComp is the benchmark that creates the most daylight between them.
For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.