Model reference — Documents
Grammar and style rewriting corrects and reshapes written text: fixing errors, shortening what is long, raising or lowering formality, and bringing wording into line with a house style guide.
The important constraint is restraint. A rewriting model must change the wording without changing the meaning, and in regulated writing it must not touch defined terms, figures or disclaimers at all. That is achieved by instructing the model narrowly and by showing the result as a marked-up difference rather than a replacement, so an author sees exactly what changed and approves it. Correction and wholesale rewriting are different jobs and are usually run as separate steps.
Models in this group take text or whole documents as input: plain text, PDFs, scanned pages and office files. Each specification table gives three hardware tiers — Minimum, the smallest setup on which the model runs correctly; Medium, the usual production configuration; and High, a configuration sized for peak volume. The rate is the number of sample inputs processed per hour on that hardware. Use these rates for initial sizing. Before production, benchmark your own data to validate accuracy, latency, throughput and cost. The sample input here is one A4 page of about 500 words corrected and rewritten.
Vendor: Meta
What it does: rewrites to a described style while preserving meaning, and follows a house style guide supplied as instructions. The choice for text that will be published.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | 2× RTX 4090 (reduced precision) | H100 80 GB | 4× H100 80 GB |
| VRAM | 48 GB combined | 80 GB | 320 GB combined |
| vCPUs | 16 | 24 | 64 |
| RAM | 64 GB | 128 GB | 512 GB |
| Server | 2× RTX 4090 24 GB | 1× H100 SXM 80 GB | 4× H100 SXM 80 GB |
| Rate (pages/hour) | ≈ 250 | ≈ 1,100 | ≈ 4,500 |
Vendor: Alibaba Cloud
What it does: rewrites long documents in one pass, keeping terminology and tone consistent from beginning to end rather than drifting between sections.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | RTX 4090 (reduced precision) | L40S 48 GB | 2× H100 80 GB |
| VRAM | 22 GB | 48 GB | 160 GB combined |
| vCPUs | 12 | 16 | 48 |
| RAM | 48 GB | 64 GB | 256 GB |
| Server | 1× RTX 4090 24 GB | 1× L40S 48 GB | 2× H100 SXM 80 GB |
| Rate (pages/hour) | ≈ 350 | ≈ 1,100 | ≈ 4,200 |
Vendor: Mistral AI
What it does: a compact model for high-volume correction — internal email, ticket replies, product copy — at low cost on a single card.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | RTX 4090 | L40S 48 GB | H100 80 GB |
| VRAM | 24 GB | 48 GB | 80 GB |
| vCPUs | 12 | 16 | 32 |
| RAM | 48 GB | 64 GB | 128 GB |
| Server | 1× RTX 4090 24 GB | 1× L40S 48 GB | 1× H100 SXM 80 GB |
| Rate (pages/hour) | ≈ 700 | ≈ 2,000 | ≈ 6,000 |
Vendor: Google
What it does: makes conservative edits that stay close to the author’s wording, which suits regulated writing where a rewrite is a risk in itself.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | RTX 4090 (reduced precision) | L40S 48 GB | 2× H100 80 GB |
| VRAM | 22 GB | 48 GB | 160 GB combined |
| vCPUs | 12 | 16 | 48 |
| RAM | 48 GB | 64 GB | 256 GB |
| Server | 1× RTX 4090 24 GB | 1× L40S 48 GB | 2× H100 SXM 80 GB |
| Rate (pages/hour) | ≈ 500 | ≈ 1,500 | ≈ 5,200 |
Vendor: Microsoft
What it does: a small model that is reliable on mechanical correction — grammar, agreement, punctuation — without attempting to improve the prose. Cheap enough to run on everything written.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | RTX 3090 | RTX 4090 | H100 80 GB |
| VRAM | 16 GB | 24 GB | 80 GB |
| vCPUs | 8 | 12 | 24 |
| RAM | 32 GB | 48 GB | 128 GB |
| Server | 1× RTX 3090 24 GB | 1× RTX 4090 24 GB | 1× H100 SXM 80 GB |
| Rate (pages/hour) | ≈ 1,000 | ≈ 3,000 | ≈ 8,000 |
Vendor: Microsoft
What it does: small purpose-built editing models that take one instruction at a time — "make this shorter", "make this formal" — and apply it narrowly. Predictable and fast, which suits an editing feature inside an application.
| Requirement | Minimum | Medium | High |
|---|---|---|---|
| GPU type | RTX 3060 | RTX 4090 | A100 80 GB |
| VRAM | 6 GB | 24 GB | 80 GB |
| vCPUs | 4 | 8 | 24 |
| RAM | 16 GB | 32 GB | 64 GB |
| Server | 1× RTX 3060 12 GB | 1× RTX 4090 24 GB | 1× A100 SXM 80 GB |
| Rate (pages/hour) | ≈ 3,000 | ≈ 12,000 | ≈ 34,000 |
The choice depends on how much latitude the model should have and whether output is reviewed. Our consultants review your style guide and a sample of your writing, then recommend a model, the instructions that hold it to your rules, and the review presentation that keeps authors in control.
At the start of a project we may run a short proof of concept on a sample of your own data, measuring the accuracy and the throughput the model actually achieves on your material. That replaces the estimates on this page with real figures, so the cost and the schedule for the full engagement are known before it is committed.
Share a representative sample, expected volume, latency target and deployment location for Grammar / style rewriting. We benchmark the shortlisted models, recommend the lowest-cost GPU configuration that meets the target, and scale it from pilot capacity to a dedicated production cluster.