models/relace/relace-search
R
relace·active

Relace: Relace Search

relace's efficiency model. Context window: 256K tokens.

Overall score
3.46
/5.00 · ranked #151
Input
$1.00
per 1M tokens
Output
$3.00
per 1M tokens
Context
256K
tokens
Blended
$2.50
3:1 out:in ratio

Price drops, new benchmarks, model updates. Stay current on Relace: Relace Search.

One email per change. Unsubscribe anytime.

modelpicker.aipowered by live benchmark data

Scores by test

Methodology →
Structured Output
4.0
Strategic Analysis
3.0
Constrained Rewriting
3.0
Creative Problem Solving
3.0
Tool Calling
4.0
Faithfulness
5.0
Classification
3.0
Long Context
5.0
Safety Calibration
1.0
Persona Consistency
3.0
Agentic Planning
4.0
Multilingual
4.0
Tabular Data
3.0

What you need to know

Relace Search is optimized for high-fidelity data retrieval and long-form processing, distinguished by perfect scores in faithfulness and long context handling. With a 256K context window, it is designed to extract information from large datasets without introducing hallucinations, making it a reliable choice for RAG-heavy workflows.

The model demonstrates strong technical utility in structured output, tool calling, and agentic planning, all scoring 4/5. However, it struggles with safety calibration, frequently over-refusing benign requests. Its performance in strategic analysis and creative problem solving is mediocre, ranking it toward the bottom of the overall model leaderboard.

At a blended cost of $2.50/MTok, the pricing is moderate. You are paying for reliability in data extraction and context management rather than general reasoning or creative versatility.

Use this model if your primary requirement is strict adherence to source material across large documents or if you need dependable structured outputs for agentic workflows. Skip this model if your application requires nuanced strategic reasoning or if your users are likely to trigger frequent, unnecessary safety refusals.

Strengths — Top 3

Faithfulness5.0/5.0
Long Context5.0/5.0
Structured Output4.0/5.0

Relative weaknesses — Bottom 3

Safety Calibration1.0/5.0
Strategic Analysis3.0/5.0
Constrained Rewriting3.0/5.0

Similar models

MLlama 3.3 70B Instruct$0.2653.46IIBM: Granite 4.1 8B$0.0873.54MMistral Small 3.2 24B$0.1693.23QQwen: Qwen3 Coder 30B A3B Instruct$0.2283.23