Head-to-head comparison across 3benchmark categories. Overall scores shown here use BenchLM's provisional ranking lane.
GPT-5.4
87
Sakana Fugu Ultra
100
Verified leaderboard positions: GPT-5.4 #19 · Sakana Fugu Ultra unranked
Pick Sakana Fugu Ultra if you want the stronger benchmark profile. GPT-5.4 only becomes the better choice if you want the cheaper token bill or you need the larger 1.05M context window.
Agentic
+5.1 difference
Coding
+25.8 difference
Knowledge
+29.4 difference
GPT-5.4
Sakana Fugu Ultra
$2.5 / $15
$5 / $30
74 t/s
N/A
151.79s
N/A
1.05M
1M
Pick Sakana Fugu Ultra if you want the stronger benchmark profile. GPT-5.4 only becomes the better choice if you want the cheaper token bill or you need the larger 1.05M context window.
Sakana Fugu Ultra is clearly ahead on the provisional aggregate, 100 to 87. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Sakana Fugu Ultra's sharpest advantage is in knowledge, where it averages 95.5 against 66.1. The single biggest benchmark swing on the page is SWE-bench Pro, 57.7% to 73.7%.
Sakana Fugu Ultra is also the more expensive model on tokens at $5.00 input / $30.00 output per 1M tokens, versus $2.50 input / $15.00 output per 1M tokens for GPT-5.4. That is roughly 2.0x on output cost alone. GPT-5.4 gives you the larger context window at 1.05M, compared with 1M for Sakana Fugu Ultra.
Sakana Fugu Ultra is ahead on BenchLM's provisional leaderboard, 100 to 87. The biggest single separator in this matchup is SWE-bench Pro, where the scores are 57.7% and 73.7%.
Sakana Fugu Ultra has the edge for knowledge tasks in this comparison, averaging 95.5 versus 66.1. Inside this category, GPQA is the benchmark that creates the most daylight between them.
Sakana Fugu Ultra has the edge for coding in this comparison, averaging 83.5 versus 57.7. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.
Sakana Fugu Ultra has the edge for agentic tasks in this comparison, averaging 82.1 versus 77. Inside this category, Terminal-Bench 2.0 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.