Best LLMs for Content Set Relevance Scoring

Category: Relevance, Classification & Matching · Typical I/O: 43546→5191 tokens

Runs on model group: All models. Inputs may reach any model in the group, not only the ranked models below.

Run this task on Fronset — request an invitation

Models

Frontier on this task: Xiaomi MiMo V2.6 Pro at 6.97 / 10. Quality bar at 90%: 6.28.

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 does not clear the bar.

Model	Quality score	CI low	Cost / 1k runs	vs best value
DeepSeek V4.1 Flash	6.94 / 10	6.53	$2.55	best value
Xiaomi MiMo V2.6 Pro	6.97 / 10	6.59	$3.60	1.4x more expensive
GPT-6 Luna	6.30 / 10	5.81	$3.77	1.5x more expensive
Thinking Machines Inkling Small	6.64 / 10	6.29	$13.31	5.2x more expensive
GPT-6 Sol	6.75 / 10	6.28	$70.68	28x more expensive
MiniMax M3	5.46 / 10	5.18	$11.13	4.4x more expensive
Gemini 3.5 Flash Lite	5.72 / 10	5.23	$7.51	2.9x more expensive
NVIDIA Nemotron-3 Nano 30B-A3B	2.50 / 10	2.14	$4.23	1.7x more expensive
NVIDIA Nemotron-3 Super 120B	5.22 / 10	4.73	$9.07	3.5x more expensive
Qwen 3.7 Plus	5.69 / 10	5.25	$15.62	6.1x more expensive
Tencent Hy3	5.84 / 10	5.40	$9.32	3.6x more expensive
Gemini 3.8 Flash	5.62 / 10	5.18	$20.77	8.1x more expensive
Claude Opus 5.5	6.15 / 10	5.73	$40.05	16x more expensive

Cost breakdown

Model	Quality	Confidence	Cost / 1k runs	Overpay	Mode
DeepSeek V4.1 Flash ★ DeepSeek	6.94 / 10 CI [6.53, 7.35]	MEDIUM	$2.55	best value	batch
Xiaomi MiMo V2.6 Pro best Xiaomi MiMo	6.97 / 10 CI [6.59, 7.36]	MEDIUM	$3.60	1.4x	batch
GPT-6 Luna OpenAI	6.30 / 10 CI [5.81, 6.79]	MEDIUM	$3.77	1.5x	batch
Thinking Machines Inkling Small OpenRouter	6.64 / 10 CI [6.29, 6.99]	MEDIUM	$13.31	5.2x	batch
GPT-6 Sol OpenAI	6.75 / 10 CI [6.28, 7.21]	MEDIUM	$70.68	28x	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: 43546 input tokens → 5191 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 complete one-to-one coverage of the supplied set, item-level semantic relevance, score calibration and consistency across items, threshold correctness, evidence quality, treatment of partial relevance, preservation of item identity or order, and resistance to keyword/source-prestige shortcuts.

Output schema

Every answer on this task is checked against this JSON Schema, whichever model wrote it. An answer that doesn't fit counts as a model failure, and the call is retried on another model.

{
  "$defs": {
    "PostRelevance": {
      "description": "Individual post relevance assessment.",
      "properties": {
        "post_id": {
          "description": "The ID or index of the post being assessed",
          "title": "Post Id",
          "type": "string"
        },
        "relevance_score": {
          "description": "Relevance score from 0.0 to 1.0",
          "maximum": 1,
          "minimum": 0,
          "title": "Relevance Score",
          "type": "number"
        }
      },
      "required": [
        "post_id",
        "relevance_score"
      ],
      "title": "PostRelevance",
      "type": "object"
    }
  },
  "description": "Analysis of post relevance for a batch of posts.",
  "properties": {
    "irrelevant_posts": {
      "description": "List of irrelevant posts with their scores",
      "items": {
        "$ref": "#/$defs/PostRelevance"
      },
      "title": "Irrelevant Posts",
      "type": "array"
    },
    "relevant_posts": {
      "description": "List of relevant posts with their scores",
      "items": {
        "$ref": "#/$defs/PostRelevance"
      },
      "title": "Relevant Posts",
      "type": "array"
    }
  },
  "required": [
    "relevant_posts",
    "irrelevant_posts"
  ],
  "title": "PostRelevanceAnalysis",
  "type": "object"
}

Prompt templates

The published template for this task.

default · LLMB_CONTENT_SET_RELEVANCE_SCORING_SYSTEM + LLMB_CONTENT_SET_RELEVANCE_SCORING_USER (4044 calls in window)

System prompt

Evaluate every item in the supplied content set for substantive relevance, not mere keyword overlap. Consider whether each item contains information that can advance the supplied query or objective, and distinguish direct evidence, useful context, passing mention, and irrelevance. Score each item independently based only on its supplied content; do not infer quality from source prestige alone. Apply the caller’s threshold consistently and return one result for every input item. 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 — content_type: {content_type}; query: {query}; posts: {posts}; relevance_threshold: {relevance_threshold}.

Relevance scale — score every item on exactly these levels, and report relevance_score as the chosen level divided by the highest level:
{relevance_levels}

Use only these inputs to complete the task defined by the system prompt.