Performance Benchmarks
Real-world performance across models and edit methods
Time Performance Analysis
Total time including generation and apply (seconds)
18.9s
Fastest Morph Time
55.2s
Average Morph Time
49%
Average Speed Improvement
Detailed Results
Click column headers to sort
Model ↕ | Method | Turns ↕Average number of tool calls used to complete the edit. | Total Time ↕Sum of generation and apply time | Apply Time ↕Time to apply the generated edit to the file | Success Rate ↕Accuracy judged by an LLM comparing the edit output against the expected diff |
|---|---|---|---|---|---|
claude-opus-4 | Morph | 2.05 | 110.5s | 1464ms | 100.0%✓ |
claude-opus-4 | Search & Replace | 5.92 | 193.9s | 0ms | 95.0% |
claude-opus-4.1 | Morph | 2.17 | 51.3s | 1586ms | 100.0%✓ |
claude-opus-4.1 | Search & Replace | 6.10 | 141.9s | 0ms | 97.5% |
claude-sonnet-4 | Morph | 2.12 | 18.9s | 1621ms | 100.0%✓ |
claude-sonnet-4 | Search & Replace | 5.78 | 32.7s | 0ms | 95.0% |
deepseek-chat | Morph | 2.12 | 44.8s | 1944ms | 100.0%✓ |
deepseek-chat | Search & Replace | 7.38 | 167.1s | 0ms | 88.5% |
gpt-5 | Morph | 2.05 | 95.2s | 1444ms | 100.0%✓ |
gpt-5 | Search & Replace | 2.55 | 124.7s | 0ms | 100.0%✓ |
grok-code-fast-1 | Morph | 2.15 | 19.6s | 2176ms | 100.0%✓ |
grok-code-fast-1 | Search & Replace | 3.69 | 35.0s | 0ms | 84.6% |
o3 | Morph | 2.12 | 45.8s | 1794ms | 100.0%✓ |
o3 | Search & Replace | 3.58 | 96.8s | 0ms | 92.5% |
Morph
Search & Replace
Cost Savings Calculator
Estimate your monthly savings with Morph vs traditional approaches
Based on Claude 4 Sonnet pricing from benchmark data
1M1B
100M tokens/day
3.0B/month
$81,912
Claude + Morph Monthly Cost
$165,153
Search & Replace
Save $83,241 (50%)
$495,460
Full Rewrite Cost
Save $413,547 (83%)
vs Search & Replace
$83,241
50% monthly savings
vs Full Rewrites
$413,547
83% monthly savings
Calculation assumptions▼
• Full rewrite cost estimated at 3× baseline method cost
• Size multipliers: Small (0.5×), Medium (1×), Large (2.5×)
• Calculations include model inference costs only
• Based on real benchmark performance data
Build the best coding agents today
Join 500+ teams using Morph to reduce token costs and apply edits at lightning speed.
10,500+
tokens per second
The fastest way to apply LLM edits
Better • Faster • Cheaper Code Edits