Coding · Benchmark profile
A harder competitive-programming benchmark family built from Codeforces, ICPC, and IOI problems, with quarter-specific public leaderboards and difficulty-aware reporting.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on LiveCodeBench Pro
Sakana Fugu-Ultra leads at 90.8%, followed by Sakana Fugu (87.8%) and GPT-5.4 (87.5%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
8 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| Sakana Fugu-Ultra | #1 | Sakana AI | 90.8% |
| Sakana Fugu | #2 | Sakana AI | 87.8% |
| GPT-5.4 | #3 | OpenAI | 87.5% |
| Gemini 3.1 Pro | #4 | 82.9% | |
| Muse Spark | #5 | Meta | 80% |
| Grok 4.20 | #6 | xAI | 74.2% |
| Claude Opus 4.6 | #7 | Anthropic | 70.7% |
| MiniCPM5-1B | #8 | OpenBMB | 22.7% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Sakana Fugu-Ultra sakana-fugu-ultra | Sakana AI | closed | Estimated reference | 90.8% |
| #2 | Sakana Fugu sakana-fugu | Sakana AI | closed | Estimated reference | 87.8% |
| #3 | GPT-5.4 gpt-5.4 | OpenAI | closed | Reference only | 87.5% |
| #4 | Gemini 3.1 Pro gemini-3-1-pro | closed | Estimated reference | 82.9% | |
| #5 | Muse Spark muse-spark | Meta | closed | Reference only | 80% |
| #6 | Grok 4.20 grok-4.20 | xAI | closed | Estimated reference | 74.2% |
| #7 | Claude Opus 4.6 claude-opus-4-6 | Anthropic | closed | Reference only | 70.7% |
| #8 | MiniCPM5-1B minicpm5-1b | OpenBMB | open | Estimated reference | 22.7% |
The top of this snapshot is led by Sakana Fugu-Ultra at 90.8%; third place is 3.3 points behind. The top-8 spread is 68.1 points.
About LiveCodeBench Pro
A harder competitive-programming benchmark family built from Codeforces, ICPC, and IOI problems, with quarter-specific public leaderboards and difficulty-aware reporting. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A harder competitive-programming benchmark family built from Codeforces, ICPC, and IOI problems, with quarter-specific public leaderboards and difficulty-aware reporting.
Sakana Fugu-Ultra by Sakana AI currently leads with 90.8%.
8 models in the LuminaBench cohort have a qualifying score on this benchmark.
Related
