Personal Coding Evals (core10): leaderboard
One engineer's 10 coding tasks taken from his daily work, each run three times per harness, model and effort configuration (Pi, Codex, Claude Code, DeepSeek Harness); the score is the share of the 30 runs solved. It is a personal, saturated suite (16 of 33 admitted configurations solve every run), not a general coding benchmark.
Metric: pass@1: mean solve rate over 10 personal coding tasks x 3 runs (%). Source: ibragim.dev. Saturation forecast: Estimated already saturated. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (Low) | 100 | #6 (GPT-6 Astra) |
| 2 | GPT-6.1 Sol (High) | 100 | #10 (GPT-6.1 Sol) |
| 3 | GPT-6.1 Sol (Medium) | 100 | #10 (GPT-6.1 Sol) |
| 4 | Claude Opus 5.5 (High) | 100 | #5 (Claude Opus 5.5) |
| 5 | Claude Haiku 5.5 (High) | 100 | #65 (Claude Haiku 5.5) |
| 6 | GPT-6 Luna (xHigh) | 100 | #95 (GPT-6 Luna) |
| 7 | GPT-6 Sol (Medium) | 100 | #25 (GPT-6 Sol) |
| 8 | DeepSeek V4.1 Flash (1017) (High) | 100 | #39 (DeepSeek V4.1 Flash (1017)) |
| 9 | Claude Opus 5.5 (Medium) | 100 | #5 (Claude Opus 5.5) |
| 10 | Claude Sonnet 5.5 (High) | 100 | #19 (Claude Sonnet 5.5) |
| 11 | Claude Opus 5.5 (Low) | 100 | #5 (Claude Opus 5.5) |
| 12 | GLM-5.3 Flash (Max) | 96.7 | #122 (GLM-5.3 Flash) |
| 13 | Claude Opus 5 (High) | 96.7 | #22 (Claude Opus 5) |
| 14 | GPT-6 Luna (Medium) | 96.7 | #95 (GPT-6 Luna) |
| 15 | GPT-5.6 Sol (Medium) | 96.7 | #30 (GPT-5.6 Sol) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=personal-coding-evals-core10 · How It Works · Data refreshed daily, snapshot 2026-11-13.