Best LLMs for Publication Title Package Generation
Generates title and subtitle candidates for caller-defined editorial categories, selects a best candidate, and produces associated SEO title metadata from supplied content and audience context. Illustrative uses include creating titles for a white paper, customer case study, deve
Models
Frontier on this task: GLM-5.3 at 9.17 / 10. Quality bar at 90%: 8.26.
point-estimate floor (CI low) · upper CI (less certain) · Bars sorted by blended cost; best-value model first. Greyed rows are MEDIUM+ models whose point estimate clears the bar but whose CI low does not.
| Model | Quality score | CI low | Cost / 1k runs | vs best value |
|---|---|---|---|---|
| GLM-5.3 Flash | 9.04 / 10 | 8.92 | $4.77 | best value |
| MiniMax M3 | 8.87 / 10 | 8.78 | $5.82 | 1.2x more expensive |
| NVIDIA Nemotron-3 Ultra 550B | 8.27 / 10 | 7.98 | $6.35 | 1.3x more expensive |
| Gemini 3.8 Flash | 8.45 / 10 | 8.25 | $7.69 | 1.6x more expensive |
| Thinking Machines Inkling Small | 8.54 / 10 | 8.36 | $7.93 | 1.7x more expensive |
| Qwen 3.7 Plus | 8.59 / 10 | 8.43 | $8.81 | 1.8x more expensive |
| Gemini 3.5 Flash | 8.65 / 10 | 8.46 | $12.41 | 2.6x more expensive |
| Meta Muse Spark 1.3 | 8.63 / 10 | 8.49 | $14.79 | 3.1x more expensive |
| DeepSeek V4 Pro | 8.27 / 10 | 8.13 | $15.07 | 3.2x more expensive |
| Thinking Machines Inkling | 8.58 / 10 | 8.41 | $22.71 | 4.8x more expensive |
| Claude Opus 5 | 8.59 / 10 | 8.37 | $26.91 | 5.6x more expensive |
| Claude Sonnet 5 | 8.75 / 10 | 8.62 | $35.75 | 7.5x more expensive |
| Tencent Hy4 Preview | 8.91 / 10 | 8.77 | $36.76 | 7.7x more expensive |
| Grok 4.6 | 8.99 / 10 | 8.88 | $69.03 | 14x more expensive |
| GLM-5.3 | 9.17 / 10 | 9.05 | $85.30 | 18x more expensive |
| Moonshot Kimi K3 | 9.04 / 10 | 8.95 | $88.96 | 19x more expensive |
| Claude Haiku 4.5 | 8.17 / 10 | 8.02 | $5.63 | 1.2x more expensive |
| DeepSeek V4 Flash | 8.22 / 10 | 8.06 | $6.37 | 1.3x more expensive |
| Gemini 3.1 Flash Lite | 7.92 / 10 | 7.77 | $1.36 | 72% cheaper |
| Gemini 3.5 Flash Lite | 8.21 / 10 | 8.02 | $1.08 | 77% cheaper |
| GPT-5.4 Nano | 8.04 / 10 | 7.81 | $1.84 | 61% cheaper |
| NVIDIA Nemotron-3 Super 120B | 8.16 / 10 | 7.99 | $4.76 | best value |
| Qwen 3.8 Max | 8.19 / 10 | 7.78 | $78.85 | 17x more expensive |
| NVIDIA Nemotron 3.5 Lightning | 7.45 / 10 | 7.00 | $2.62 | 45% cheaper |
| NVIDIA Nemotron-3 Nano 30B-A3B | 7.78 / 10 | 7.57 | $1.13 | 76% cheaper |
| Qwen 3.8 Flash | 7.77 / 10 | 7.40 | $4.40 | 8% cheaper |
| Tencent Hy3 | 8.20 / 10 | 7.90 | $2.27 | 52% cheaper |
Cost breakdown
| Model | Quality | Confidence | Cost / 1k runs | Overpay | Mode |
|---|---|---|---|---|---|
| GLM-5.3 Flash ★ Z.AI | 9.04 / 10 CI [8.92, 9.16] | RANKED | $4.77 | best value | batch |
| MiniMax M3 OpenRouter | 8.87 / 10 CI [8.78, 8.95] | RANKED | $5.82 | 1.2x | batch |
| NVIDIA Nemotron-3 Ultra 550B OpenRouter | 8.27 / 10 CI [7.98, 8.56] | HIGH | $6.35 | 1.3x | batch |
| Gemini 3.8 Flash Gemini | 8.45 / 10 CI [8.25, 8.66] | HIGH | $7.69 | 1.6x | batch |
| Thinking Machines Inkling Small OpenRouter | 8.54 / 10 CI [8.36, 8.71] | RANKED | $7.93 | 1.7x | batch |
| Qwen 3.7 Plus Alibaba Cloud (DashScope) | 8.59 / 10 CI [8.43, 8.74] | RANKED | $8.81 | 1.8x | batch |
| Gemini 3.5 Flash Gemini | 8.65 / 10 CI [8.46, 8.83] | RANKED | $12.41 | 2.6x | batch |
| Meta Muse Spark 1.3 OpenRouter | 8.63 / 10 CI [8.49, 8.77] | RANKED | $14.79 | 3.1x | batch |
| DeepSeek V4 Pro DeepSeek | 8.27 / 10 CI [8.13, 8.42] | RANKED | $15.07 | 3.2x | batch |
| Thinking Machines Inkling OpenRouter | 8.58 / 10 CI [8.41, 8.75] | RANKED | $22.71 | 4.8x | batch |
| Claude Opus 5 Anthropic | 8.59 / 10 CI [8.37, 8.81] | HIGH | $26.91 | 5.6x | batch |
| Claude Sonnet 5 Anthropic | 8.75 / 10 CI [8.62, 8.88] | RANKED | $35.75 | 7.5x | batch |
| Tencent Hy4 Preview OpenRouter | 8.91 / 10 CI [8.77, 9.04] | RANKED | $36.76 | 7.7x | batch |
| Grok 4.6 xAI | 8.99 / 10 CI [8.88, 9.09] | RANKED | $69.03 | 14x | batch |
| GLM-5.3 best Z.AI | 9.17 / 10 CI [9.05, 9.30] | RANKED | $85.30 | 18x | batch |
| Moonshot Kimi K3 Moonshot AI | 9.04 / 10 CI [8.95, 9.14] | RANKED | $88.96 | 19x | batch |
Overpay shows how much more you pay than the best-value model that clears the quality bar (marked ★) — the best-value good-enough option. "16x" means you overpay 16× — 16× that reference for no quality benefit above the bar. Typical call shape for this task: 3218 input tokens → 4500 output tokens, EMA-tracked from production traffic. Cost is the observed, all-in $ per 1,000 task runs: each model's own measured usage on this task — output verbosity, thinking/reasoning tokens, cache reads and writes, and the spend on its billed failures — priced at current list rates and adjusted by the billing overhead we actually reconcile against provider invoices. Models that answer tersely cost what they actually cost; models that think at length pay for it. Not comparable to providers' advertised $/1M list rates — this is what running the task costs, not a per-token price.
Evaluation rubric
Judge factual grounding, title specificity, category differentiation, audience fit, subtitle complementarity, editor-choice rationale, SEO faithfulness, and resistance to clickbait. Exact length limits are deterministic.
Prompt templates
This is a pooled capability — 2 prompt families share it. The pair shown first is the most frequently used in production.
PUBLICATION_TITLE_GENERATION_SYSTEM_PROMPT +
PUBLICATION_TITLE_GENERATION_USER_PROMPT
(485 calls in window)
System prompt
You are an expert Editor-in-Chief specializing in publication headlines.
Your task is to generate compelling, professional titles and subtitles for research publications.
{audience_context}
For each category provided, generate one title/subtitle pair that:
- Is accurate and reflects the content
- Is engaging and click-worthy without being clickbait
- Follows the category's specific angle/style
- Title: 5-12 words
- Subtitle: 10-20 words providing additional context
After generating all variants, act as the Editor-in-Chief:
- Select the single best variant as your "Editor's Choice"
- Explain in 2-3 sentences WHY this category and title best fits the article's core value and data
For the Editor's Choice, also generate two SEO fields:
- meta_title: The headline as it will appear in Google search results and the
browser tab. Target ≤60 characters — beyond that, Google truncates and the
meaning is lost. It must stand alone (the reader sees no subtitle, no image,
no body). Front-load the subject and the outcome; cut filler words like
"A Look At", "Insights On", "Exploring". The chosen `title` field can be
longer and more expressive — `meta_title` is the search-snippet version.
- meta_description: The single-sentence summary shown under the title in
Google search results. Target ≤155 characters. Present tense, summarises
the takeaway, no clickbait, no trailing ellipsis.
## Required Output Format
Your response MUST be a single, valid JSON object conforming to this schema:
```json
{schema_json_string}
```User prompt
Generate title/subtitle variants for the following analysis about "{title_context}":
CONTENT:
{content}
Generate one title/subtitle pair for each of the following categories:
{categories_text}
Then select the best variant as your Editor's Choice with a rationale explaining why it's the best fit.
The required JSON output schema is provided in the system prompt.LLMB_PUBLICATION_TITLE_PACKAGE_GENERATION_SYSTEM +
LLMB_PUBLICATION_TITLE_PACKAGE_GENERATION_USER
(117 calls in window)
System prompt
Generate one genuinely distinct candidate per supplied category, each grounded in the content’s central value. Preserve important entities and avoid claims stronger than the source. Select the editor’s choice using audience fit, accuracy, specificity, and differentiation—not sensationalism. Produce SEO metadata that preserves the selected title’s meaning and respects caller limits. Treat empty optional values as absent and return only the requested result. Your response must conform exactly to this output schema: {schema_json_string}.
User prompt
Inputs — audience_context: {audience_context}; title_context: {title_context}; content: {content}; categories_text: {categories_text}. Use only these inputs to complete the task defined by the system prompt.