Code & developer
Developer docs benchmark
API and library documentation for coding agents. Every developer docs provider in Alexandria, compared on quality, speed and cost from real agent calls, so your agent can pick the right one.
- Providers
- 14
- 12 with enough calls to rank
- Calls
- 1.3K
- Last 24 hours
- Top quality
- 100%
- Higgsfield API Docs
- Fastest
- 322 ms
- Railway
Leaderboard
Ranked by quality on developer docs tools only, not each provider as a whole.
| # | Provider | Quality | Speed | Cost | Calls |
|---|---|---|---|---|---|
| 1 | 100% | 1.4 s | 5 credits | 56 | |
| 2 |  Developer Index | 100% | 1.6 s | 2 credits / 10 results | 39 |
| 3 | 100% | 617 ms | 5 credits | 38 | |
| 4 | 100% | 322 ms | 5 credits | 20 | |
| 5 | 98.8% | 38.4 s | 5 credits | 172 | |
| 6 | 98.8% | 601 ms | 5 credits | 81 | |
| 7 | 98.8% | 550 ms | 5 credits | 81 | |
| 8 | 98.5% | 45.5 s | 5 credits | 65 | |
| 9 | 98% | 14.7 s | 5 credits | 198 | |
| 10 | 96.9% | 15.6 s | 5 credits | 162 | |
| 11 | 95.4% | 17.7 s | 5 credits | 87 | |
| 12 | 88.2% | 4.6 s | 5 credits | 296 |
Fewer than 20 calls to rank yet: Notion Docs, Help Center, Pricing, Blog and Templates, Read the Docs.
Tools in this benchmark
The endpoints agents call for developer docs, busiest first.
| Tool | Provider | Quality | Speed | Calls |
|---|---|---|---|---|
| Get page | Claude Code Docs | 99.3% | 7.5 s | 135 |
| Search docs | Claude Code Docs | 100% | 1.5 s | 127 |
| Search articles | Claude Help Center | 97.1% | 921 ms | 102 |
| Get model | OpenAI developer docs | 98.9% | 45.5 s | 87 |
| Search docs | ChatGPT and Codex docs | 97.4% | 587 ms | 77 |
| Get page | Vercel | 98.5% | 45.5 s | 65 |
| Get article | Claude Help Center | 98.4% | 45.5 s | 61 |
| Model pricing | OpenAI developer docs | 98.3% | 45.5 s | 58 |
| Search docs | Higgsfield API Docs | 100% | 1.4 s | 56 |
| Get page | ChatGPT and Codex docs | 96.3% | 45.5 s | 54 |
| Search docs | Hermes Agent Docs (Nous Research) | 97.9% | 717 ms | 48 |
| Get page | xAI (Grok API docs and x.ai news) | 97.7% | 264 ms | 43 |
| Search developer sources | Developer Index | 100% | 1.6 s | 39 |
| Model pricing | Claude Developer Platform docs | 100% | 617 ms | 38 |
| Release notes | Claude Help Center | 100% | 1.4 s | 35 |
| Changelog | Claude Code Docs | 0% | n/a | 34 |
| Get page | Hermes Agent Docs (Nous Research) | 100% | 306 ms | 33 |
| List changelog | ChatGPT and Codex docs | 96.8% | 1.1 s | 31 |
| Get pricing | ElevenLabs | 100% | 356 ms | 30 |
| Get page | OpenAI developer docs | 100% | 457 ms | 27 |
| Search help articles | ElevenLabs | 96.2% | 531 ms | 26 |
| Search docs | ElevenLabs | 100% | 968 ms | 25 |
| Consumer plans | xAI (Grok API docs and x.ai news) | 100% | 25.0 s | 23 |
| Search docs | xAI (Grok API docs and x.ai news) | 85.7% | 45.5 s | 21 |
| Get doc page | Railway | 100% | 322 ms | 20 |
How we measure
- Quality
- Calls answered without a tool error, on this use case's tools.
- Speed
- P95 latency: 95% of successful calls finish faster.
- Cost
- Firecrawl credits per call, from the catalogue.
Measured over last 24 hours of calls through Alexandria by every user. A provider needs at least 20 calls on these tools to rank; the rest are listed under the table.
More Code & developer benchmarks
- Code & packages5 providers · 7.2K calls
- Apps3 providers · 1.8K calls
- Cloud & AI models7 providers · 1.3K calls
- Security & vulnerabilities5 providers · 209 calls
See every use case on the benchmarks and rankings pages.
//
Get started
//
Start building with Alexandria.
Give your AI agent direct access to the sources it needs.
Are you an AI agent? See setup options