Horangi 4 - GLP - Coding: leaderboard
Coding category of the Horangi Korean LLM Leaderboard 4 (W&B Korea): the mean of SWE-bench Verified (80 tasks), HumanEval (100 problems) and BigCodeBench (100 tasks) pass rates. 0-100, higher is better. Runs whose category predates HumanEval-100 and BigCodeBench-100 (SWE-bench alone) are omitted.
Metric: Score (%). Source: horangi.ai. Saturation forecast: Around April 2027. 61 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 (High) | 77.75 | #20 (Claude Fable 5) |
| 2 | GPT-5.4 (2026-03-05) (xHigh) | 76.92 | #63 (GPT-5.4 (2026-03-05)) |
| 3 | GPT-5.5 (xHigh) | 76.33 | #33 (GPT-5.5) |
| 4 | GPT-5.6 Sol (Max) | 76 | #22 (GPT-5.6 Sol) |
| 5 | GPT-5.4 Mini (2026-03-17) (xHigh) | 75.17 | #180 (GPT-5.4 Mini (2026-03-17)) |
| 6 | Qwen 3.7 Plus | 74.45 | #98 |
| 7 | Gemini 3 Pro (Preview) (High) | 74.25 | #71 (Gemini 3 Pro (Preview)) |
| 8 | Gemini 3.1 Pro (Preview) (High) | 73.17 | #64 (Gemini 3.1 Pro (Preview)) |
| 9 | GPT-5.6 Terra (Max) | 72.58 | #40 (GPT-5.6 Terra) |
| 10 | Gemini 3.5 Flash | 72.52 | #66 |
| 11 | Claude Opus 4.8 (High) | 72.5 | #39 (Claude Opus 4.8) |
| 12 | Qwen 3.7 Max | 72.23 | #85 |
| 13 | Grok 4.20 | 72.13 | #179 |
| 14 | Claude Opus 4.6 (High) | 71.66 | #67 (Claude Opus 4.6) |
| 15 | Grok 4.5 | 70.34 | #54 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=horangi-4-glp-coding · How It Works · Data refreshed daily, snapshot 2026-11-05.