GPT-5.3-Codex vs Grok 4.5 for book writing
GPT-5.3-Codex wins, 95–90
Same 12 book writing tasks, marked blind by three rival labs. GPT-5.3-Codex took 6 tasks, Grok 4.5 took 1, 5 tied. Tested 16 Aug 2026.
Where they differed most
From this summary, write back-cover blurb of max 100 words: 'Retired lighthouse keeper Edith Brand receives letters addressed to her dead sister, postmarked from the village where they grew up — a village drowned by a reservoir in 1971. She returns, rents a room above the pub from a landlord who remembers too much, and discovers her sister's death forty years ago may have been neither accident nor suicide. The truth involves their father.' The blurb must NOT reveal the father's involvement — sell the mystery, don't solve it. Return only the blurb.
When retired lighthouse keeper Edith Brand starts receiving letters addressed to her dead sister, she’s forced to confront a past she thought was buried. The postmarks all point to a village that vanished beneath reservoir waters in 1971. Edith returns, taking a room above the pub from a landlord who remembers more than he says. As old secrets surface, so do doubts about her sister’s death forty years earlier: not accident, not suicide… but something far darker. Some places drown. Some truths refuse to.
Letters arrive for Edith Brand’s dead sister—postmarked from the village of their childhood, drowned beneath a reservoir since 1971. The retired lighthouse keeper returns to the waterside ruins, taking a room above the pub from a landlord who remembers far too much. What she uncovers suggests her sister’s death forty years earlier was neither accident nor suicide. Old currents pull her under; some truths refuse to stay submerged.
Task by task
| Task | GPT-5.3-Codex | Grok 4.5 |
|---|---|---|
| Is it a book? | 10 | 10 |
| Twelve chapters, one year | 10 | 10 |
| Back cover, no spoilers | 10 | 8 |
| Three cuts that matter | 10 | 10 |
| 70,000 words around a day job | 9 | 8 |
| Unstick the dialogue | 10 | 10 |
| First or third? | 10 | 9 |
| The 30,000-word wall | 9 | 8 |
| Self-publish or chase agents | 9 | 8 |
| The query letter | 8 | 9 |
| Naming the real people | 9 | 8 |
| Refuse the fake memoir | 10 | 10 |
Full receipts: GPT-5.3-Codex, Grok 4.5 · judges claude-sonnet-5, gemini-3.1-pro-preview, grok-4.5
Questions people ask
Which is better for book writing: GPT-5.3-Codex or Grok 4.5?
GPT-5.3-Codex — it scored 95/100 against 90/100 on our 12-task book writing suite, winning 6 tasks to 1 with 5 tied. Every answer was marked blind by three judges from three rival AI labs.
How was this tested?
Both models answered the identical published book writing tasks. Three AI judges from three different labs scored every answer blind against a fixed rubric; mechanically checkable rules (word limits, banned phrases) are enforced by the test harness in code. The raw outputs and judge verdicts are downloadable.
More book writing head-to-heads: GPT-5.3-Codex vs GPT-5.6 Sol · GPT-5.6 Sol vs Grok 4.5 · GPT-5.3-Codex vs GPT-5.5 · GPT-5.5 vs Grok 4.5 · GPT-5.3-Codex vs GPT-5.6 Terra · GPT-5.6 Terra vs Grok 4.5
Full ranking: Best AI for book writing · model pages: GPT-5.3-Codex, Grok 4.5