Relace
2 models built · 2 served
Relace (Squack Inc., founded 2022, US) builds specialized models for AI coding agents rather than general-purpose chat models. Its focus is narrow and infrastructure-level: Fast Apply merge models that take a code edit and apply it to a file at 4,300+ tokens/sec, plus retrieval models for locating relevant code. On OpenKey, Relace has two models: Relace Apply 3, for merging edits into existing code, and Relace Search, for finding the right code context. Both use a 256K context window. This is a lab to consider if you're building a coding agent pipeline and need a dedicated apply/search step rather than a chat model doing everything.
Lineup
With only two models, the choice is about function, not budget. Relace Apply 3 ($0.85/M input, $1.25/M output) is the merge model — use it for applying diffs or edits to existing files fast. Relace Search ($1.00/M input, $3.00/M output) is the retrieval model — use it to locate relevant code spans before editing. Most coding-agent setups will want both: Search to find the target, Apply 3 to make the change. Neither is a general chat model, so don't route unrelated tasks to them.
- Relace SearchDec 8, 2025 256K $1.03 $3.09 ctx 256K · in $1.03 · out $3.09
- Relace Apply 3Sep 26, 2025 256K $0.875 $1.29 ctx 256K · in $0.875 · out $1.29
Prices per 1M tokens, flat 3% fee included.
Models Relace serves
Beyond its own models, Relace hosts 2 models in the OpenKey catalog. Prices are Relace’s own list price per 1M input tokens, before our fee.
| Model | Built by | Context | Input /M |
|---|---|---|---|
| Relace Apply 3 | Relace | 256K | $0.850 |
| Relace Search | Relace | 256K | $1.00 |
From the OpenRouter catalog, Jul 28, 2026. All providers
Privacy policy ↗Terms of service ↗
Questions
- What does Relace pricing look like on OpenKey?
- Relace Apply 3 lists at $0.85/M input and $1.25/M output tokens; on OpenKey that's $0.8755/M input and $1.2875/M output (provider price × 1.03). Relace Search lists at $1.00/M input and $3.00/M output, or $1.03/M input and $3.09/M output on OpenKey. Both use a 256K context window.
- Which Relace model should I use?
- It depends on the task, not price — the two models do different jobs. Use Relace Apply 3 to merge code edits into files (that's the Fast Apply model running at 4,300+ tokens/sec). Use Relace Search to retrieve relevant code before editing. A coding agent typically needs both in sequence.
- Does Relace offer free models?
- No. Relace has 2 models on OpenKey and neither is free. Both are priced per-token: Relace Apply 3 at $0.85/$1.25 per M input/output tokens, and Relace Search at $1.00/$3.00 per M input/output tokens, plus OpenKey's flat 3% fee on top of list price.