Benchmarks · Human preference
Khalil Bench
Not a serious benchmark: what one X user, @kamkmgamer, replied to Adam Holter's model test posts, each reply mapped to a verdict from yuck to really cool and scored 0 to 100. Most models have one or two verdicts. Never in a default mix and given no quality score.
21 models · 21 results · updated 2026-11-13 · khalil-bench.adamholter.chatgpt.site ↗
Leaderboard
Fable 5.1
More from this benchmark
4 published measuresScore vs any published measure4 measures: sub-scores, splits, counts
8 more models under All models in settings.
Breakdown
darker is better within each column · click a column to sort| Model | ||||
|---|---|---|---|---|
| Fable 5.1 | 1 | 1.5 | 1.5 | 1.5 |
| Fable 5.2 | 2 | 1.5 | 1.5 | 1.5 |
| Gemini 4 Carbon | 2 | 1.25 | 1 | 1.5 |
| Opus 5.5 | 4 | 0.5 | -1 | 1 |
| DeepSeek V4.1 Flash | 3 | 0.33 | 0 | 1 |
| MiMo V2.6 Pro | 3 | 0.33 | -1 | 1 |
| Sonnet 5.5 | 3 | 0.17 | -1.5 | 1 |
| Muse Wonder | 2 | 0 | 0 | 0 |
| GPT-6.1 Terra | 2 | 0 | 0 | 0 |
| Inkling 2 Small | 1 | 0 | 0 | 0 |
| GPT-6 Luna | 4 | -0.12 | -1 | 1 |
| Gemini 4 Hydrogen | 6 | -0.17 | -1 | 1 |
| MiniMax M3.1 | 8 | -0.19 | -2 | 1 |
| GPT-6 Astra | 1 | -0.5 | -0.5 | -0.5 |
| GPT-6 Sol | 2 | -0.5 | -1 | 0 |
| Mistral Large 4 | 5 | -0.5 | -1.5 | 0 |
| Gemini 4 Argon | 2 | -1 | -2 | 0 |
| MiMo V2.6 Flash | 1 | -1 | -1 | -1 |
| Gemini 3.8 Flash | 1 | -2 | -2 | -2 |
| Grok 4.7 | 1 | -2 | -2 | -2 |
| Kimi K3.1 | 1 | -2 | -2 | -2 |
All results
21 rows| # | Model | Effort | Khalil Bench |
|---|---|---|---|
| 1 | Fable 5.1 | 88.0 | |
| 2 | Fable 5.2 | 88.0 | |
| 3 | Gemini 4 Carbon | 81.0 | |
| 4 | Opus 5.5 | 63.0 | |
| 5 | DeepSeek V4.1 Flash | 58.0 | |
| 6 | MiMo V2.6 Pro | 58.0 | |
| 7 | Sonnet 5.5 | 54.0 | |
| 8 | Muse Wonder | 50.0 | |
| 9 | GPT-6.1 Terra | 50.0 | |
| 10 | Inkling 2 Small | 50.0 | |
| 11 | GPT-6 Luna | 47.0 | |
| 12 | Gemini 4 Hydrogen | 46.0 | |
| 13 | MiniMax M3.1 | 45.0 | |
| 14 | GPT-6 Astra | 38.0 | |
| 15 | GPT-6 Sol | 38.0 | |
| 16 | Mistral Large 4 | 38.0 | |
| 17 | Gemini 4 Argon | 25.0 | |
| 18 | MiMo V2.6 Flash | 25.0 | |
| 19 | Gemini 3.8 Flash | 0.0 | |
| 20 | Grok 4.7 | 0.0 | |
| 21 | Kimi K3.1 | 0.0 |