Grok 4.5 vs Sakana Fugu-Ultra
Head-to-head comparison across 2benchmark categories. Overall scores shown here use BenchLM's provisional ranking lane.
Verdict
Sakana Fugu-Ultra leads for most workloads.
Based on BenchLM composite scores, July 2026.
Grok 4.5
75
Sakana Fugu-Ultra
80
Pick Sakana Fugu-Ultra if you want the stronger benchmark profile. Grok 4.5 only becomes the better choice if agentic is the priority.
Category Breakdown
| Benchmark | Grok 4.5 | Δ | Sakana Fugu-Ultra |
|---|---|---|---|
| Coding | 64.7 | → 4.9 | 69.6 |
| Agentic | 83.3 | ← 1.2 | 82.1 |
| Reasoning | — | — | 93.6 |
| Knowledge | — | — | 95.5 |
| Multimodal | — | — | 86.6 |
Operational Comparison
Grok 4.5
Sakana Fugu-Ultra
$2 / $6
$null / $null
N/A
N/A
N/A
N/A
500K
1M
Quick Verdict
Pick Sakana Fugu-Ultra if you want the stronger benchmark profile. Grok 4.5 only becomes the better choice if agentic is the priority.
Sakana Fugu-Ultra is clearly ahead on the provisional aggregate, 80 to 75. 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 coding, where it averages 69.6 against 64.7. The single biggest benchmark swing on the page is SWE-bench Pro, 64.7% to 73.7%. Grok 4.5 does hit back in agentic, so the answer changes if that is the part of the workload you care about most.
Sakana Fugu-Ultra gives you the larger context window at 1M, compared with 500K for Grok 4.5.
Benchmark Deep Dive
Frequently Asked Questions (3)
Which is better, Grok 4.5 or Sakana Fugu-Ultra?
Sakana Fugu-Ultra is ahead on BenchLM's provisional leaderboard, 80 to 75. The biggest single separator in this matchup is SWE-bench Pro, where the scores are 64.7% and 73.7%.
Which is better for coding, Grok 4.5 or Sakana Fugu-Ultra?
Sakana Fugu-Ultra has the edge for coding in this comparison, averaging 69.6 versus 64.7. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, Grok 4.5 or Sakana Fugu-Ultra?
Grok 4.5 has the edge for agentic tasks in this comparison, averaging 83.3 versus 82.1. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
The AI models change fast. We track them for you.
For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.