Morph
2 models built · 7 served
Morph is a small lab from AutoInfra, Inc., founded in 2023 and based in the United States. It ships fast inference models built specifically for coding agents, with Fast Apply as its calling card — a technique for merging code edits into files at roughly 10,500 tokens/sec. Rather than competing on general chat or reasoning benchmarks, Morph optimizes for one job: applying diffs and edits quickly enough to keep an agentic coding loop fast. The lineup is just two models, Morph V3 Fast and Morph V3 Large, both released July 7, 2025, both text-to-text. If your workload involves an AI coding agent that needs to patch files at speed rather than write long-form reasoning, this is the lab built for that.
Lineup
Two models, both priced for high-volume edit application. Morph V3 Fast runs $0.80/1M input and $1.20/1M output tokens, with an 81,920-token context window — pick this for routine edit-apply calls where cost per call matters and files fit comfortably in that window. Morph V3 Large costs $0.90/1M input and $1.90/1M output tokens but extends context to 262,144 tokens, useful when your agent needs to apply edits against large files or multi-file diffs. Neither model is free. Start with V3 Fast for typical agent loops; move to V3 Large only when you're hitting the context ceiling.
- Morph V3 FastJul 7, 2025 82K $0.824 $1.24 ctx 82K · in $0.824 · out $1.24
- Morph V3 LargeJul 7, 2025 262K $0.927 $1.96 ctx 262K · in $0.927 · out $1.96
Prices per 1M tokens, flat 3% fee included.
Models Morph serves
Beyond its own models, Morph hosts 7 models in the OpenKey catalog. Prices are Morph’s own list price per 1M input tokens, before our fee.
| Model | Built by | Context | Input /M |
|---|---|---|---|
| DeepSeek V4 Flash | DeepSeek | 1.0M | $0.139 |
| Gemma 4 31B | 262K | $0.140 | |
| MiniMax M3 | MiniMax | 1.0M | $0.300 |
| Morph V3 Fast | Morph | 82K | $0.800 |
| Morph V3 Large | Morph | 262K | $0.900 |
| Qwen3.6 27B | Qwen | 262K | $0.289 |
| GLM 5.2 | Z.ai | 1.0M | $1.10 |
From the OpenRouter catalog, Jul 28, 2026. All providers
Privacy policy ↗Terms of service ↗
Questions
- What does Morph's pricing look like on OpenKey?
- Morph V3 Fast is $0.80/1M input and $1.20/1M output from the provider, which becomes $0.824/1M and $1.236/1M on OpenKey after the flat 3% fee. Morph V3 Large is $0.90/1M input and $1.90/1M output, or $0.927/1M and $1.957/1M on OpenKey. Both are inexpensive relative to general-purpose flagship models.
- Which Morph model should I use, and is either free?
- Neither of Morph's two models is free. Morph V3 Fast is the cheaper option and fits an 81,920-token context — good for standard edit-apply tasks. Morph V3 Large costs a bit more per token but supports a 262,144-token context, so use it when your agent needs to apply edits across larger files or codebases.
- What is Morph actually known for?
- Morph is known for Fast Apply, a technique for merging code edits into existing files at roughly 10,500 tokens/sec. It's built by AutoInfra, Inc., a US company founded in 2023, and its whole model lineup — V3 Fast and V3 Large, both shipped July 7, 2025 — is aimed at speeding up coding agents rather than general-purpose chat or reasoning.