# API Agent Landscape

A market map for API teams: which products agents can discover, understand, and use in real tasks. Claude Code, Codex, CLI, and MCP are evidence surfaces; the primary lens is API company readiness.

[Compare the leading AI visibility and GEO tools for developer teams](/blog/best-geo-tools-for-developer-tools-2026), including Lightsage, Profound, Otterly, AthenaHQ, and Peec.

| Metric | Value |
|--------|-------|
| API companies ranked | 185 |
| Models | 2 |
| API categories covered | 44 |
| API benchmark signals | 367 |
| Average API score | 55 |

## Interpretation

For an API company, the benchmark is really asking whether agents can become successful API users: can they find the right docs, understand the product, authenticate, recover from errors, and finish a task without human intervention?

API is the primary market lens. CLI and MCP stay in the landscape because they reveal whether readiness holds across multiple entry points, not just one happy-path integration.

| Surface | Claude Code | Codex | Read |
|---------|-------------|-------|------|
| API | 55 | 56 | Even |
| CLI | 36 | 57 | Codex +21 |
| MCP | 38 | 39 | Even |

Rankings below favor companies with both Claude Code and Codex API rows when possible. One-off high scores remain useful signals, but they should not carry the same weight as repeated API performance.

## What API Teams Should Take Away

- **API readiness baseline:** 185 API companies average 55 across 367 scored API runs.
- **Interface risk:** CLI changes the market view most, with a 21-point swing.
- **Agent effort:** The fastest API slice is Codex at 2m 10s.
- **Friction hotspot:** Agent Memory averages 10 errors and 6m 50s runtime.

## Agent Lens For API Companies

| Agent | API Score | API Companies | API Categories | Avg Cost | Avg Runtime | Avg Tool Calls | Avg Errors |
|-------|-----------|---------------|----------------|----------|-------------|----------------|------------|
| Claude Code | 55 | 185 | 44 | $0.49 | 4m 4s | 18 | 4 |
| Codex | 56 | 182 | 43 | $0.35 | 2m 13s | 23 | 4 |

## Claude Vs Codex Surface Map

| Interface | Agent | Average Score | Companies | Categories | Current Leader |
|-----------|-------|---------------|-----------|------------|----------------|
| API | Claude Code | 55 | 185 | 44 | [OpenRouter](/leaderboard/claudecode/api/openrouter) |
| API | Codex | 56 | 185 | 44 | [ElevenLabs](/leaderboard/codex/api/elevenlabs) |
| CLI | Claude Code | 36 | 98 | 24 | [Deepgram](/leaderboard/claudecode/cli/deepgram) |
| CLI | Codex | 57 | 61 | 21 | [LiveKit](/leaderboard/codex/cli/livekit) |
| MCP | Claude Code | 38 | 80 | 26 | [fastCRW](/leaderboard/claudecode/mcp/fastcrw) |
| MCP | Codex | 39 | 79 | 26 | [Jina AI](/leaderboard/codex/mcp/jina-ai) |

## Category Leaders By Agent And Surface

These tables mirror the default rendered page state: one leader per category and surface.

### Claude Code Category Leaders

| Category | API | CLI | MCP |
|----------|-----|-----|-----|
| Inference | [OpenRouter](/leaderboard/claudecode/api/openrouter) | [OpenRouter](/leaderboard/claudecode/cli/openrouter) | [Cerebras](/leaderboard/claudecode/mcp/cerebras) |
| Payment | [Stripe](/leaderboard/claudecode/api/stripe) | [Stripe](/leaderboard/claudecode/cli/stripe) | [Stripe](/leaderboard/claudecode/mcp/stripe) |
| Graph Databases | [Neo4j](/leaderboard/claudecode/api/neo4j) | — | — |
| Voice STT | [Deepgram](/leaderboard/claudecode/api/deepgram) | [Deepgram](/leaderboard/claudecode/cli/deepgram) | [Deepgram](/leaderboard/claudecode/mcp/deepgram) |
| Voice TTS | [ElevenLabs](/leaderboard/claudecode/api/elevenlabs) | [Rime](/leaderboard/claudecode/cli/rime) | [ElevenLabs](/leaderboard/claudecode/mcp/elevenlabs) |
| Sandboxes | [Superserve](/leaderboard/claudecode/api/superserve) | [Tenki](/leaderboard/claudecode/cli/tenki) | [E2B](/leaderboard/claudecode/mcp/e2b) |
| Vector Databases | [Chroma](/leaderboard/claudecode/api/chroma) | [Pinecone](/leaderboard/claudecode/cli/pinecone) | [Qdrant](/leaderboard/claudecode/mcp/qdrant) |
| Auth | [WorkOS](/leaderboard/claudecode/api/workos) | [WorkOS](/leaderboard/claudecode/cli/workos) | [Descope](/leaderboard/claudecode/mcp/descope) |
| Document Parsing | [Reducto](/leaderboard/claudecode/api/reducto) | [Extend.ai](/leaderboard/claudecode/cli/extend-ai) | [Extend.ai](/leaderboard/claudecode/mcp/extend-ai) |
| Browser | [Browserbase](/leaderboard/claudecode/api/browserbase) | [Steel](/leaderboard/claudecode/cli/steel) | [Anchor Browser](/leaderboard/claudecode/mcp/anchor-browser) |
| Durable Workflow | [Prefect](/leaderboard/claudecode/api/prefect) | [Temporal](/leaderboard/claudecode/cli/temporal) | [Prefect](/leaderboard/claudecode/mcp/prefect) |
| Search | [Firecrawl](/leaderboard/claudecode/api/firecrawl) | [Tavily](/leaderboard/claudecode/cli/tavily) | [Jina AI](/leaderboard/claudecode/mcp/jina-ai) |
| Observability | [Datadog](/leaderboard/claudecode/api/datadog) | [Datadog](/leaderboard/claudecode/cli/datadog) | [Datadog](/leaderboard/claudecode/mcp/datadog) |
| Storage | [Box](/leaderboard/claudecode/api/box) | [Box](/leaderboard/claudecode/cli/box) | [Box](/leaderboard/claudecode/mcp/box) |
| Cloud Hosting | [Render](/leaderboard/claudecode/api/render) | [Cloudflare](/leaderboard/claudecode/cli/cloudflare) | [Render](/leaderboard/claudecode/mcp/render) |
| Audio | [Sonilo](/leaderboard/claudecode/api/sonilo) | — | — |
| Geocoding | [OpenCageData](/leaderboard/claudecode/api/opencagedata) | — | [OpenCageData](/leaderboard/claudecode/mcp/opencagedata) |
| Meeting Bot | [Meetstream](/leaderboard/claudecode/api/meetstream) | [MeetGeek](/leaderboard/claudecode/cli/meetgeek) | [MeetGeek](/leaderboard/claudecode/mcp/meetgeek) |
| Webscraping | [Scrapfly](/leaderboard/claudecode/api/scrapfly) | [fastCRW](/leaderboard/claudecode/cli/fastcrw) | [fastCRW](/leaderboard/claudecode/mcp/fastcrw) |
| E-Signature | [Docusign](/leaderboard/claudecode/api/docusign) | — | — |
| Voice Telephony | [Telnyx](/leaderboard/claudecode/api/telnyx) | [Vonage](/leaderboard/claudecode/cli/vonage) | [Plivo](/leaderboard/claudecode/mcp/plivo) |
| API Testing | [Postman](/leaderboard/claudecode/api/postman) | — | — |
| Video Agent | [Tavus](/leaderboard/claudecode/api/tavus) | — | — |
| Email | [Agentmail](/leaderboard/claudecode/api/agentmail) | [Resend](/leaderboard/claudecode/cli/resend) | [Agentmail](/leaderboard/claudecode/mcp/agentmail) |
| Agent Memory | [Zep](/leaderboard/claudecode/api/zep) | — | — |
| Unified API | [Naive](/leaderboard/claudecode/api/naive) | [Naive](/leaderboard/claudecode/cli/naive) | [Naive](/leaderboard/claudecode/mcp/naive) |
| Video Infra | [Mux](/leaderboard/claudecode/api/mux) | — | — |
| Proxy | [Massive](/leaderboard/claudecode/api/massive) | — | [Massive](/leaderboard/claudecode/mcp/massive) |
| Media Generation | [Replicate](/leaderboard/claudecode/api/replicate) | — | — |
| Voice Infra | [Agora](/leaderboard/claudecode/api/agora) | [LiveKit](/leaderboard/claudecode/cli/livekit) | [LiveKit](/leaderboard/claudecode/mcp/livekit) |
| Stablecoin | [Coinbase Payments](/leaderboard/claudecode/api/coinbase-payments) | [Circle](/leaderboard/claudecode/cli/circle) | [BlindPay](/leaderboard/claudecode/mcp/blindpay) |
| Coding Agent | [Kosuke](/leaderboard/claudecode/api/kosuke) | — | — |
| Verification | [Prelude](/leaderboard/claudecode/api/prelude) | — | — |
| Agent Automation | [Dust](/leaderboard/claudecode/api/dust) | — | — |
| Agent UI | [CopilotKit](/leaderboard/claudecode/api/copilotkit) | — | — |
| Sales Intelligence | [Apollo](/leaderboard/claudecode/api/apollo) | [Apollo](/leaderboard/claudecode/cli/apollo) | [Apollo](/leaderboard/claudecode/mcp/apollo) |
| Public Sector Intelligence | [GovSpend](/leaderboard/claudecode/api/govspend) | — | — |
| CMS | [Shopware](/leaderboard/claudecode/api/shopware) | — | — |
| Data Pipeline | [CocoIndex](/leaderboard/claudecode/api/cocoindex) | — | — |
| Database | [Convex](/leaderboard/claudecode/api/convex) | — | [Convex](/leaderboard/claudecode/mcp/convex) |
| Neocloud | [Nebius AI Cloud](/leaderboard/claudecode/api/nebius-ai-cloud) | [Nebius AI Cloud](/leaderboard/claudecode/cli/nebius-ai-cloud) | — |
| Code Review | [Macroscope](/leaderboard/claudecode/api/macroscope) | [Semgrep](/leaderboard/claudecode/cli/semgrep) | — |
| Talent Directory | [Roster](/leaderboard/claudecode/api/roster) | — | — |
| Security Scanning | [Unclone](/leaderboard/claudecode/api/unclone) | — | — |
| Agent MCP Gateway | — | — | [Composio](/leaderboard/claudecode/mcp/composio) |

### Codex Category Leaders

| Category | API | CLI | MCP |
|----------|-----|-----|-----|
| Voice TTS | [ElevenLabs](/leaderboard/codex/api/elevenlabs) | [ElevenLabs](/leaderboard/codex/cli/elevenlabs) | [ElevenLabs](/leaderboard/codex/mcp/elevenlabs) |
| Sandboxes | [Superserve](/leaderboard/codex/api/superserve) | [Tenki](/leaderboard/codex/cli/tenki) | [Daytona](/leaderboard/codex/mcp/daytona) |
| Voice STT | [Deepgram](/leaderboard/codex/api/deepgram) | [Deepgram](/leaderboard/codex/cli/deepgram) | [Deepgram](/leaderboard/codex/mcp/deepgram) |
| Webscraping | [fastCRW](/leaderboard/codex/api/fastcrw) | [fastCRW](/leaderboard/codex/cli/fastcrw) | [fastCRW](/leaderboard/codex/mcp/fastcrw) |
| Audio | [Sonilo](/leaderboard/codex/api/sonilo) | — | — |
| Document Parsing | [Reducto](/leaderboard/codex/api/reducto) | [LlamaParse](/leaderboard/codex/cli/llamaparse) | [Extend.ai](/leaderboard/codex/mcp/extend-ai) |
| Inference | [Fireworks AI](/leaderboard/codex/api/fireworks-ai) | [OpenRouter](/leaderboard/codex/cli/openrouter) | [Cerebras](/leaderboard/codex/mcp/cerebras) |
| Browser | [Browser Use](/leaderboard/codex/api/browser-use) | [Browserbase](/leaderboard/codex/cli/browserbase) | [Anchor Browser](/leaderboard/codex/mcp/anchor-browser) |
| Observability | [Datadog](/leaderboard/codex/api/datadog) | [Datadog](/leaderboard/codex/cli/datadog) | [Datadog](/leaderboard/codex/mcp/datadog) |
| Vector Databases | [Qdrant](/leaderboard/codex/api/qdrant) | [Pinecone](/leaderboard/codex/cli/pinecone) | [Pinecone](/leaderboard/codex/mcp/pinecone) |
| Auth | [WorkOS](/leaderboard/codex/api/workos) | [WorkOS](/leaderboard/codex/cli/workos) | [Descope](/leaderboard/codex/mcp/descope) |
| Durable Workflow | [Restate](/leaderboard/codex/api/restate) | [Camunda](/leaderboard/codex/cli/camunda) | [Prefect](/leaderboard/codex/mcp/prefect) |
| Graph Databases | [Neo4j](/leaderboard/codex/api/neo4j) | — | — |
| Search | [Firecrawl](/leaderboard/codex/api/firecrawl) | [Firecrawl](/leaderboard/codex/cli/firecrawl) | [Jina AI](/leaderboard/codex/mcp/jina-ai) |
| Email | [Agentmail](/leaderboard/codex/api/agentmail) | [Resend](/leaderboard/codex/cli/resend) | [Agentmail](/leaderboard/codex/mcp/agentmail) |
| Payment | [Stripe](/leaderboard/codex/api/stripe) | [Stripe](/leaderboard/codex/cli/stripe) | [Stripe](/leaderboard/codex/mcp/stripe) |
| Meeting Bot | [Meetstream](/leaderboard/codex/api/meetstream) | — | [MeetGeek](/leaderboard/codex/mcp/meetgeek) |
| Video Agent | [HeyGen](/leaderboard/codex/api/heygen) | — | — |
| Video Infra | [Mux](/leaderboard/codex/api/mux) | — | — |
| Voice Telephony | [Vapi](/leaderboard/codex/api/vapi) | [Vonage](/leaderboard/codex/cli/vonage) | [Twilio](/leaderboard/codex/mcp/twilio) |
| Cloud Hosting | [Render](/leaderboard/codex/api/render) | [Cloudflare](/leaderboard/codex/cli/cloudflare) | [Render](/leaderboard/codex/mcp/render) |
| Coding Agent | [Kosuke](/leaderboard/codex/api/kosuke) | — | — |
| Agent Memory | [Letta](/leaderboard/codex/api/letta) | — | — |
| API Testing | [Postman](/leaderboard/codex/api/postman) | — | — |
| Geocoding | [OpenCageData](/leaderboard/codex/api/opencagedata) | — | [OpenCageData](/leaderboard/codex/mcp/opencagedata) |
| Unified API | [Apideck](/leaderboard/codex/api/apideck) | [Naive](/leaderboard/codex/cli/naive) | [Naive](/leaderboard/codex/mcp/naive) |
| Voice Infra | [LiveKit](/leaderboard/codex/api/livekit) | [LiveKit](/leaderboard/codex/cli/livekit) | [LiveKit](/leaderboard/codex/mcp/livekit) |
| Storage | [Dropbox](/leaderboard/codex/api/dropbox) | [Box](/leaderboard/codex/cli/box) | [Box](/leaderboard/codex/mcp/box) |
| Proxy | [Massive](/leaderboard/codex/api/massive) | — | [Massive](/leaderboard/codex/mcp/massive) |
| Media Generation | [Replicate](/leaderboard/codex/api/replicate) | — | — |
| Agent UI | [CopilotKit](/leaderboard/codex/api/copilotkit) | — | — |
| E-Signature | [BoldSign](/leaderboard/codex/api/boldsign) | — | — |
| Verification | [Surge](/leaderboard/codex/api/surge) | — | — |
| Stablecoin | [DFNS](/leaderboard/codex/api/dfns) | — | [BlindPay](/leaderboard/codex/mcp/blindpay) |
| Sales Intelligence | [Apollo](/leaderboard/codex/api/apollo) | — | [Apollo](/leaderboard/codex/mcp/apollo) |
| Agent Automation | [Dust](/leaderboard/codex/api/dust) | — | — |
| Public Sector Intelligence | [GovSpend](/leaderboard/codex/api/govspend) | — | — |
| CMS | [Shopware](/leaderboard/codex/api/shopware) | — | — |
| Data Pipeline | [CocoIndex](/leaderboard/codex/api/cocoindex) | — | — |
| Neocloud | [Nebius AI Cloud](/leaderboard/codex/api/nebius-ai-cloud) | [Nebius AI Cloud](/leaderboard/codex/cli/nebius-ai-cloud) | — |
| Code Review | [Macroscope](/leaderboard/codex/api/macroscope) | [Semgrep](/leaderboard/codex/cli/semgrep) | — |
| Talent Directory | [Roster](/leaderboard/codex/api/roster) | — | — |
| Security Scanning | [Unclone](/leaderboard/codex/api/unclone) | — | — |
| Database | — | — | [Convex](/leaderboard/codex/mcp/convex) |
| Agent MCP Gateway | — | — | [Composio](/leaderboard/codex/mcp/composio) |

## API, CLI, And MCP Chart Sections

The rendered Agent Landscape page ends with API, CLI, and MCP chart sections. Each section starts with the same overall chart used on the public arena pages, then repeats that chart by category. The model toggle switches Claude Code and Codex. Charts with no evaluated companies are omitted; in charts with evaluated companies, skipped rows remain visible as skipped rows.

| Surface | Agent | Rows | Scored | Skipped | Categories | Average Score | Avg Runtime | Avg Cost |
|---------|-------|------|--------|---------|------------|---------------|-------------|----------|
| API | Claude Code | 185 | 121 | 64 | 44 | 71 | 4m 7s | $0.49 |
| API | Codex | 185 | 121 | 64 | 44 | 71 | 2m 13s | $0.35 |
| CLI | Claude Code | 98 | 50 | 48 | 24 | 64 | 2m 48s | $0.27 |
| CLI | Codex | 61 | 52 | 9 | 21 | 62 | 2m 14s | $0.28 |
| MCP | Claude Code | 80 | 49 | 31 | 26 | 54 | 5m 24s | $0.31 |
| MCP | Codex | 79 | 50 | 29 | 26 | 54 | 3m 40s | $0.17 |

### Category Chart Data

Each row below corresponds to one category chart in the API, CLI, or MCP chart sections.

| Surface | Agent | Category | Rows | Scored | Skipped | Average Score | Current Leader |
|---------|-------|----------|------|--------|---------|---------------|----------------|
| API | Claude Code | Inference | 14 | 7 | 7 | 76 | [OpenRouter](/leaderboard/claudecode/api/openrouter) |
| API | Claude Code | Payment | 12 | 7 | 5 | 71 | [Stripe](/leaderboard/claudecode/api/stripe) |
| API | Claude Code | Sandboxes | 11 | 8 | 3 | 80 | [Superserve](/leaderboard/claudecode/api/superserve) |
| API | Claude Code | Search | 10 | 8 | 2 | 74 | [Firecrawl](/leaderboard/claudecode/api/firecrawl) |
| API | Claude Code | Voice Telephony | 10 | 5 | 5 | 68 | [Telnyx](/leaderboard/claudecode/api/telnyx) |
| API | Claude Code | Durable Workflow | 8 | 2 | 6 | 75 | [Prefect](/leaderboard/claudecode/api/prefect) |
| API | Claude Code | Stablecoin | 8 | 4 | 4 | 60 | [Coinbase Payments](/leaderboard/claudecode/api/coinbase-payments) |
| API | Claude Code | Cloud Hosting | 7 | 5 | 2 | 63 | [Render](/leaderboard/claudecode/api/render) |
| API | Claude Code | Audio | 6 | 2 | 4 | 69 | [Sonilo](/leaderboard/claudecode/api/sonilo) |
| API | Claude Code | Auth | 6 | 6 | 0 | 75 | [WorkOS](/leaderboard/claudecode/api/workos) |
| API | Claude Code | Vector Databases | 6 | 5 | 1 | 75 | [Chroma](/leaderboard/claudecode/api/chroma) |
| API | Claude Code | Voice STT | 6 | 4 | 2 | 76 | [Deepgram](/leaderboard/claudecode/api/deepgram) |
| API | Claude Code | Voice TTS | 6 | 5 | 1 | 76 | [ElevenLabs](/leaderboard/claudecode/api/elevenlabs) |
| API | Claude Code | Document Parsing | 5 | 5 | 0 | 73 | [Reducto](/leaderboard/claudecode/api/reducto) |
| API | Claude Code | Meeting Bot | 5 | 4 | 1 | 66 | [Meetstream](/leaderboard/claudecode/api/meetstream) |
| API | Claude Code | Voice Infra | 5 | 3 | 2 | 65 | [Agora](/leaderboard/claudecode/api/agora) |
| API | Claude Code | Agent Memory | 4 | 4 | 0 | 69 | [Zep](/leaderboard/claudecode/api/zep) |
| API | Claude Code | Browser | 4 | 4 | 0 | 78 | [Browserbase](/leaderboard/claudecode/api/browserbase) |
| API | Claude Code | Graph Databases | 4 | 3 | 1 | 74 | [Neo4j](/leaderboard/claudecode/api/neo4j) |
| API | Claude Code | Public Sector Intelligence | 4 | 1 | 3 | 22 | [NationGraph](/leaderboard/claudecode/api/nationgraph) |
| API | Claude Code | Unified API | 4 | 2 | 2 | 71 | [Apideck](/leaderboard/claudecode/api/apideck) |
| API | Claude Code | Agent Automation | 3 | 1 | 2 | 65 | [Dust](/leaderboard/claudecode/api/dust) |
| API | Claude Code | E-Signature | 3 | 3 | 0 | 71 | [Docusign](/leaderboard/claudecode/api/docusign) |
| API | Claude Code | Email | 3 | 3 | 0 | 71 | [Agentmail](/leaderboard/claudecode/api/agentmail) |
| API | Claude Code | Observability | 3 | 2 | 1 | 71 | [Datadog](/leaderboard/claudecode/api/datadog) |
| API | Claude Code | Verification | 3 | 2 | 1 | 63 | [Prelude](/leaderboard/claudecode/api/prelude) |
| API | Claude Code | Video Infra | 3 | 3 | 0 | 68 | [Mux](/leaderboard/claudecode/api/mux) |
| API | Claude Code | Webscraping | 3 | 3 | 0 | 76 | [Scrapfly](/leaderboard/claudecode/api/scrapfly) |
| API | Claude Code | Storage | 2 | 2 | 0 | 77 | [Box](/leaderboard/claudecode/api/box) |
| API | Claude Code | Video Agent | 2 | 2 | 0 | 73 | [Tavus](/leaderboard/claudecode/api/tavus) |
| API | Claude Code | Agent UI | 1 | 1 | 0 | 64 | [CopilotKit](/leaderboard/claudecode/api/copilotkit) |
| API | Claude Code | API Testing | 1 | 1 | 0 | 74 | [Postman](/leaderboard/claudecode/api/postman) |
| API | Claude Code | Coding Agent | 1 | 1 | 0 | 66 | [Kosuke](/leaderboard/claudecode/api/kosuke) |
| API | Claude Code | Geocoding | 1 | 1 | 0 | 79 | [OpenCageData](/leaderboard/claudecode/api/opencagedata) |
| API | Claude Code | Media Generation | 1 | 1 | 0 | 69 | [Replicate](/leaderboard/claudecode/api/replicate) |
| API | Claude Code | Sales Intelligence | 1 | 1 | 0 | 52 | [Apollo](/leaderboard/claudecode/api/apollo) |
| API | Codex | Inference | 14 | 7 | 7 | 74 | [Fireworks AI](/leaderboard/codex/api/fireworks-ai) |
| API | Codex | Payment | 12 | 7 | 5 | 68 | [Stripe](/leaderboard/codex/api/stripe) |
| API | Codex | Sandboxes | 11 | 8 | 3 | 79 | [Superserve](/leaderboard/codex/api/superserve) |
| API | Codex | Search | 10 | 8 | 2 | 74 | [Firecrawl](/leaderboard/codex/api/firecrawl) |
| API | Codex | Voice Telephony | 10 | 5 | 5 | 64 | [Vapi](/leaderboard/codex/api/vapi) |
| API | Codex | Durable Workflow | 8 | 2 | 6 | 76 | [Restate](/leaderboard/codex/api/restate) |
| API | Codex | Stablecoin | 8 | 4 | 4 | 52 | [DFNS](/leaderboard/codex/api/dfns) |
| API | Codex | Cloud Hosting | 7 | 5 | 2 | 62 | [Render](/leaderboard/codex/api/render) |
| API | Codex | Audio | 6 | 2 | 4 | 69 | [Sonilo](/leaderboard/codex/api/sonilo) |
| API | Codex | Auth | 6 | 6 | 0 | 77 | [WorkOS](/leaderboard/codex/api/workos) |
| API | Codex | Vector Databases | 6 | 5 | 1 | 76 | [Qdrant](/leaderboard/codex/api/qdrant) |
| API | Codex | Voice STT | 6 | 4 | 2 | 77 | [Deepgram](/leaderboard/codex/api/deepgram) |
| API | Codex | Voice TTS | 6 | 5 | 1 | 78 | [ElevenLabs](/leaderboard/codex/api/elevenlabs) |
| API | Codex | Document Parsing | 5 | 5 | 0 | 73 | [Reducto](/leaderboard/codex/api/reducto) |
| API | Codex | Meeting Bot | 5 | 4 | 1 | 66 | [Meetstream](/leaderboard/codex/api/meetstream) |
| API | Codex | Voice Infra | 5 | 3 | 2 | 68 | [LiveKit](/leaderboard/codex/api/livekit) |
| API | Codex | Agent Memory | 4 | 4 | 0 | 72 | [Letta](/leaderboard/codex/api/letta) |
| API | Codex | Browser | 4 | 4 | 0 | 76 | [Browser Use](/leaderboard/codex/api/browser-use) |
| API | Codex | Graph Databases | 4 | 3 | 1 | 69 | [Neo4j](/leaderboard/codex/api/neo4j) |
| API | Codex | Public Sector Intelligence | 4 | 1 | 3 | 30 | [NationGraph](/leaderboard/codex/api/nationgraph) |
| API | Codex | Unified API | 4 | 2 | 2 | 69 | [Apideck](/leaderboard/codex/api/apideck) |
| API | Codex | Agent Automation | 3 | 1 | 2 | 52 | [Dust](/leaderboard/codex/api/dust) |
| API | Codex | E-Signature | 3 | 3 | 0 | 60 | [BoldSign](/leaderboard/codex/api/boldsign) |
| API | Codex | Email | 3 | 3 | 0 | 74 | [Agentmail](/leaderboard/codex/api/agentmail) |
| API | Codex | Observability | 3 | 2 | 1 | 75 | [Datadog](/leaderboard/codex/api/datadog) |
| API | Codex | Verification | 3 | 2 | 1 | 66 | [Surge](/leaderboard/codex/api/surge) |
| API | Codex | Video Infra | 3 | 3 | 0 | 71 | [Mux](/leaderboard/codex/api/mux) |
| API | Codex | Webscraping | 3 | 3 | 0 | 76 | [fastCRW](/leaderboard/codex/api/fastcrw) |
| API | Codex | Storage | 2 | 2 | 0 | 70 | [Dropbox](/leaderboard/codex/api/dropbox) |
| API | Codex | Video Agent | 2 | 2 | 0 | 73 | [HeyGen](/leaderboard/codex/api/heygen) |
| API | Codex | Agent UI | 1 | 1 | 0 | 68 | [CopilotKit](/leaderboard/codex/api/copilotkit) |
| API | Codex | API Testing | 1 | 1 | 0 | 73 | [Postman](/leaderboard/codex/api/postman) |
| API | Codex | Coding Agent | 1 | 1 | 0 | 74 | [Kosuke](/leaderboard/codex/api/kosuke) |
| API | Codex | Geocoding | 1 | 1 | 0 | 72 | [OpenCageData](/leaderboard/codex/api/opencagedata) |
| API | Codex | Media Generation | 1 | 1 | 0 | 69 | [Replicate](/leaderboard/codex/api/replicate) |
| API | Codex | Sales Intelligence | 1 | 1 | 0 | 57 | [Apollo](/leaderboard/codex/api/apollo) |
| CLI | Claude Code | Payment | 9 | 1 | 8 | 80 | [Stripe](/leaderboard/claudecode/cli/stripe) |
| CLI | Claude Code | Inference | 8 | 3 | 5 | 41 | [OpenRouter](/leaderboard/claudecode/cli/openrouter) |
| CLI | Claude Code | Durable Workflow | 7 | 2 | 5 | 61 | [Temporal](/leaderboard/claudecode/cli/temporal) |
| CLI | Claude Code | Sandboxes | 7 | 6 | 1 | 81 | [Tenki](/leaderboard/claudecode/cli/tenki) |
| CLI | Claude Code | Search | 7 | 4 | 3 | 78 | [Tavily](/leaderboard/claudecode/cli/tavily) |
| CLI | Claude Code | Auth | 6 | 4 | 2 | 60 | [WorkOS](/leaderboard/claudecode/cli/workos) |
| CLI | Claude Code | Vector Databases | 6 | 2 | 4 | 56 | [Pinecone](/leaderboard/claudecode/cli/pinecone) |
| CLI | Claude Code | Cloud Hosting | 5 | 4 | 1 | 65 | [Cloudflare](/leaderboard/claudecode/cli/cloudflare) |
| CLI | Claude Code | Browser | 4 | 3 | 1 | 75 | [Steel](/leaderboard/claudecode/cli/steel) |
| CLI | Claude Code | Code Review | 4 | 2 | 2 | 46 | [Semgrep](/leaderboard/claudecode/cli/semgrep) |
| CLI | Claude Code | Document Parsing | 4 | 4 | 0 | 60 | [Extend.ai](/leaderboard/claudecode/cli/extend-ai) |
| CLI | Claude Code | Email | 3 | 2 | 1 | 80 | [Resend](/leaderboard/claudecode/cli/resend) |
| CLI | Claude Code | Voice Infra | 3 | 1 | 2 | 87 | [LiveKit](/leaderboard/claudecode/cli/livekit) |
| CLI | Claude Code | Voice STT | 3 | 2 | 1 | 84 | [Deepgram](/leaderboard/claudecode/cli/deepgram) |
| CLI | Claude Code | Voice Telephony | 3 | 3 | 0 | 49 | [Vonage](/leaderboard/claudecode/cli/vonage) |
| CLI | Claude Code | Voice TTS | 3 | 3 | 0 | 51 | [Rime](/leaderboard/claudecode/cli/rime) |
| CLI | Claude Code | Observability | 1 | 1 | 0 | 63 | [Datadog](/leaderboard/claudecode/cli/datadog) |
| CLI | Claude Code | Sales Intelligence | 1 | 1 | 0 | 21 | [Apollo](/leaderboard/claudecode/cli/apollo) |
| CLI | Claude Code | Storage | 1 | 1 | 0 | 68 | [Box](/leaderboard/claudecode/cli/box) |
| CLI | Claude Code | Unified API | 1 | 1 | 0 | 64 | [Naive](/leaderboard/claudecode/cli/naive) |
| CLI | Codex | Sandboxes | 6 | 6 | 0 | 70 | [Tenki](/leaderboard/codex/cli/tenki) |
| CLI | Codex | Search | 6 | 5 | 1 | 69 | [Firecrawl](/leaderboard/codex/cli/firecrawl) |
| CLI | Codex | Cloud Hosting | 5 | 4 | 1 | 68 | [Cloudflare](/leaderboard/codex/cli/cloudflare) |
| CLI | Codex | Auth | 4 | 4 | 0 | 61 | [WorkOS](/leaderboard/codex/cli/workos) |
| CLI | Codex | Browser | 4 | 3 | 1 | 74 | [Browserbase](/leaderboard/codex/cli/browserbase) |
| CLI | Codex | Code Review | 4 | 2 | 2 | 51 | [Semgrep](/leaderboard/codex/cli/semgrep) |
| CLI | Codex | Document Parsing | 4 | 4 | 0 | 68 | [LlamaParse](/leaderboard/codex/cli/llamaparse) |
| CLI | Codex | Vector Databases | 4 | 2 | 2 | 48 | [Pinecone](/leaderboard/codex/cli/pinecone) |
| CLI | Codex | Email | 3 | 2 | 1 | 79 | [Resend](/leaderboard/codex/cli/resend) |
| CLI | Codex | Inference | 3 | 3 | 0 | 52 | [OpenRouter](/leaderboard/codex/cli/openrouter) |
| CLI | Codex | Voice Telephony | 3 | 3 | 0 | 41 | [Vonage](/leaderboard/codex/cli/vonage) |
| CLI | Codex | Voice TTS | 3 | 3 | 0 | 53 | [ElevenLabs](/leaderboard/codex/cli/elevenlabs) |
| CLI | Codex | Durable Workflow | 2 | 2 | 0 | 51 | [Camunda](/leaderboard/codex/cli/camunda) |
| CLI | Codex | Voice Infra | 2 | 2 | 0 | 47 | [LiveKit](/leaderboard/codex/cli/livekit) |
| CLI | Codex | Voice STT | 2 | 2 | 0 | 82 | [Deepgram](/leaderboard/codex/cli/deepgram) |
| CLI | Codex | Neocloud | 1 | 1 | 0 | 39 | [Nebius AI Cloud](/leaderboard/codex/cli/nebius-ai-cloud) |
| CLI | Codex | Observability | 1 | 1 | 0 | 65 | [Datadog](/leaderboard/codex/cli/datadog) |
| CLI | Codex | Payment | 1 | 1 | 0 | 79 | [Stripe](/leaderboard/codex/cli/stripe) |
| CLI | Codex | Storage | 1 | 1 | 0 | 66 | [Box](/leaderboard/codex/cli/box) |
| CLI | Codex | Unified API | 1 | 1 | 0 | 48 | [Naive](/leaderboard/codex/cli/naive) |
| MCP | Claude Code | Search | 8 | 7 | 1 | 74 | [Jina AI](/leaderboard/claudecode/mcp/jina-ai) |
| MCP | Claude Code | Payment | 7 | 3 | 4 | 47 | [Stripe](/leaderboard/claudecode/mcp/stripe) |
| MCP | Claude Code | Voice Telephony | 7 | 3 | 4 | 47 | [Plivo](/leaderboard/claudecode/mcp/plivo) |
| MCP | Claude Code | Auth | 6 | 3 | 3 | 67 | [Descope](/leaderboard/claudecode/mcp/descope) |
| MCP | Claude Code | Vector Databases | 6 | 5 | 1 | 54 | [Qdrant](/leaderboard/claudecode/mcp/qdrant) |
| MCP | Claude Code | Cloud Hosting | 5 | 3 | 2 | 47 | [Render](/leaderboard/claudecode/mcp/render) |
| MCP | Claude Code | Inference | 5 | 1 | 4 | 11 | [Groq](/leaderboard/claudecode/mcp/groq) |
| MCP | Claude Code | Sandboxes | 4 | 2 | 2 | 48 | [E2B](/leaderboard/claudecode/mcp/e2b) |
| MCP | Claude Code | Email | 3 | 3 | 0 | 62 | [Agentmail](/leaderboard/claudecode/mcp/agentmail) |
| MCP | Claude Code | Meeting Bot | 3 | 2 | 1 | 57 | [MeetGeek](/leaderboard/claudecode/mcp/meetgeek) |
| MCP | Claude Code | Stablecoin | 3 | 1 | 2 | 38 | [BlindPay](/leaderboard/claudecode/mcp/blindpay) |
| MCP | Claude Code | Voice TTS | 3 | 3 | 0 | 71 | [ElevenLabs](/leaderboard/claudecode/mcp/elevenlabs) |
| MCP | Claude Code | Document Parsing | 2 | 2 | 0 | 75 | [Extend.ai](/leaderboard/claudecode/mcp/extend-ai) |
| MCP | Claude Code | Durable Workflow | 2 | 1 | 1 | 44 | [Prefect](/leaderboard/claudecode/mcp/prefect) |
| MCP | Claude Code | Voice Infra | 2 | 1 | 1 | 44 | [LiveKit](/leaderboard/claudecode/mcp/livekit) |
| MCP | Claude Code | Webscraping | 2 | 2 | 0 | 60 | [fastCRW](/leaderboard/claudecode/mcp/fastcrw) |
| MCP | Claude Code | Database | 1 | 1 | 0 | 20 | [Convex](/leaderboard/claudecode/mcp/convex) |
| MCP | Claude Code | Observability | 1 | 1 | 0 | 15 | [Datadog](/leaderboard/claudecode/mcp/datadog) |
| MCP | Claude Code | Proxy | 1 | 1 | 0 | 47 | [Massive](/leaderboard/claudecode/mcp/massive) |
| MCP | Claude Code | Sales Intelligence | 1 | 1 | 0 | 11 | [Apollo](/leaderboard/claudecode/mcp/apollo) |
| MCP | Claude Code | Storage | 1 | 1 | 0 | 87 | [Box](/leaderboard/claudecode/mcp/box) |
| MCP | Claude Code | Unified API | 1 | 1 | 0 | 33 | [Naive](/leaderboard/claudecode/mcp/naive) |
| MCP | Claude Code | Voice STT | 1 | 1 | 0 | 30 | [Deepgram](/leaderboard/claudecode/mcp/deepgram) |
| MCP | Codex | Search | 8 | 7 | 1 | 66 | [Jina AI](/leaderboard/codex/mcp/jina-ai) |
| MCP | Codex | Payment | 7 | 3 | 4 | 60 | [Stripe](/leaderboard/codex/mcp/stripe) |
| MCP | Codex | Voice Telephony | 7 | 3 | 4 | 46 | [Twilio](/leaderboard/codex/mcp/twilio) |
| MCP | Codex | Vector Databases | 6 | 5 | 1 | 42 | [Pinecone](/leaderboard/codex/mcp/pinecone) |
| MCP | Codex | Auth | 5 | 3 | 2 | 53 | [Descope](/leaderboard/codex/mcp/descope) |
| MCP | Codex | Cloud Hosting | 5 | 3 | 2 | 64 | [Render](/leaderboard/codex/mcp/render) |
| MCP | Codex | Inference | 5 | 1 | 4 | 31 | [Groq](/leaderboard/codex/mcp/groq) |
| MCP | Codex | Sandboxes | 4 | 2 | 2 | 63 | [Daytona](/leaderboard/codex/mcp/daytona) |
| MCP | Codex | Email | 3 | 3 | 0 | 60 | [Agentmail](/leaderboard/codex/mcp/agentmail) |
| MCP | Codex | Meeting Bot | 3 | 2 | 1 | 46 | [MeetGeek](/leaderboard/codex/mcp/meetgeek) |
| MCP | Codex | Stablecoin | 3 | 1 | 2 | 40 | [BlindPay](/leaderboard/codex/mcp/blindpay) |
| MCP | Codex | Voice TTS | 3 | 3 | 0 | 51 | [ElevenLabs](/leaderboard/codex/mcp/elevenlabs) |
| MCP | Codex | Document Parsing | 2 | 2 | 0 | 75 | [Extend.ai](/leaderboard/codex/mcp/extend-ai) |
| MCP | Codex | Durable Workflow | 2 | 1 | 1 | 48 | [Prefect](/leaderboard/codex/mcp/prefect) |
| MCP | Codex | Voice Infra | 2 | 2 | 0 | 47 | [LiveKit](/leaderboard/codex/mcp/livekit) |
| MCP | Codex | Webscraping | 2 | 2 | 0 | 78 | [fastCRW](/leaderboard/codex/mcp/fastcrw) |
| MCP | Codex | Database | 1 | 1 | 0 | 37 | [Convex](/leaderboard/codex/mcp/convex) |
| MCP | Codex | Observability | 1 | 1 | 0 | 37 | [Datadog](/leaderboard/codex/mcp/datadog) |
| MCP | Codex | Proxy | 1 | 1 | 0 | 47 | [Massive](/leaderboard/codex/mcp/massive) |
| MCP | Codex | Sales Intelligence | 1 | 1 | 0 | 39 | [Apollo](/leaderboard/codex/mcp/apollo) |
| MCP | Codex | Storage | 1 | 1 | 0 | 83 | [Box](/leaderboard/codex/mcp/box) |
| MCP | Codex | Unified API | 1 | 1 | 0 | 38 | [Naive](/leaderboard/codex/mcp/naive) |
| MCP | Codex | Voice STT | 1 | 1 | 0 | 38 | [Deepgram](/leaderboard/codex/mcp/deepgram) |

## Leaders By API, CLI, MCP, And Overall

Overall includes companies scored by both agents across API, CLI, and MCP. API, CLI, and MCP include companies scored by both agents on that interface.

### Overall

Both agents across API, CLI, and MCP.

| Rank | Company | Category | Avg Score | Rows | Avg Runtime | Avg Errors |
|------|---------|----------|-----------|------|-------------|------------|
| 1 | [Stripe](/leaderboard/claudecode/api/stripe) | Payment | 82 | 6 | 1m 32s | 0.8 |
| 2 | [Pinecone](/leaderboard/claudecode/api/pinecone) | Vector Databases | 79 | 6 | 2m 22s | 2.0 |
| 3 | [Jina AI](/leaderboard/claudecode/api/jina-ai) | Search | 78 | 6 | 2m 35s | 0.8 |
| 4 | [Firecrawl](/leaderboard/claudecode/api/firecrawl) | Search | 78 | 6 | 2m 36s | 0.8 |
| 5 | [Render](/leaderboard/claudecode/api/render) | Cloud Hosting | 77 | 6 | 3m 44s | 1.8 |
| 6 | [Extend.ai](/leaderboard/claudecode/api/extend-ai) | Document Parsing | 76 | 6 | 2m 36s | 3.0 |

### API

Claude Code + Codex.

| Rank | Company | Category | Avg Score | Rows | Avg Runtime | Avg Errors |
|------|---------|----------|-----------|------|-------------|------------|
| 1 | [ElevenLabs](/leaderboard/claudecode/api/elevenlabs) | Voice TTS | 89 | 2 | 1m 14s | 0.5 |
| 2 | [Superserve](/leaderboard/claudecode/api/superserve) | Sandboxes | 87 | 2 | 1m 7s | 4.5 |
| 3 | [AssemblyAI](/leaderboard/claudecode/api/assemblyai) | Voice STT | 86 | 2 | 1m 16s | 0.5 |
| 4 | [Deepgram](/leaderboard/claudecode/api/deepgram) | Voice STT | 86 | 2 | 54s | 1.5 |
| 5 | [Daytona](/leaderboard/claudecode/api/daytona) | Sandboxes | 86 | 2 | 60s | 3.5 |
| 6 | [Vercel](/leaderboard/claudecode/api/vercel) | Sandboxes | 85 | 2 | 1m 5s | 2.0 |

### CLI

Claude Code + Codex.

| Rank | Company | Category | Avg Score | Rows | Avg Runtime | Avg Errors |
|------|---------|----------|-----------|------|-------------|------------|
| 1 | [Deepgram](/leaderboard/claudecode/api/deepgram) | Voice STT | 90 | 2 | 1m 46s | 0.0 |
| 2 | [LiveKit](/leaderboard/claudecode/api/livekit) | Voice Infra | 89 | 2 | 1m 18s | 0.0 |
| 3 | [Tenki](/leaderboard/claudecode/api/tenki) | Sandboxes | 87 | 2 | 1m 1s | 0.0 |
| 4 | [Resend](/leaderboard/claudecode/api/resend) | Email | 84 | 2 | 1m 14s | 4.0 |
| 5 | [Firecrawl](/leaderboard/claudecode/api/firecrawl) | Search | 81 | 2 | 1m 29s | 0.0 |
| 6 | [Sprites](/leaderboard/claudecode/api/sprites) | Sandboxes | 81 | 2 | 1m 1s | 0.0 |

### MCP

Claude Code + Codex.

| Rank | Company | Category | Avg Score | Rows | Avg Runtime | Avg Errors |
|------|---------|----------|-----------|------|-------------|------------|
| 1 | [fastCRW](/leaderboard/claudecode/api/fastcrw) | Webscraping | 87 | 2 | 2m 2s | 0.0 |
| 2 | [Jina AI](/leaderboard/claudecode/api/jina-ai) | Search | 86 | 2 | 2m 37s | 0.0 |
| 3 | [Box](/leaderboard/claudecode/api/box) | Storage | 85 | 2 | 3m 36s | 0.0 |
| 4 | [Pinecone](/leaderboard/claudecode/api/pinecone) | Vector Databases | 84 | 2 | 3m 12s | 0.0 |
| 5 | [Stripe](/leaderboard/claudecode/api/stripe) | Payment | 83 | 2 | 2m 10s | 0.0 |
| 6 | [Exa](/leaderboard/claudecode/api/exa) | Search | 80 | 2 | 2m 52s | 1.0 |

## API Company Leaderboard

Top API companies by blended Claude Code and Codex API benchmark score.

| Rank | Company | Category | Score | Rows |
|------|---------|----------|-------|------|
| 1 | [ElevenLabs](/leaderboard/claudecode/api/elevenlabs) | Voice TTS | 89 | 2 |
| 2 | [Superserve](/leaderboard/claudecode/api/superserve) | Sandboxes | 87 | 2 |
| 3 | [Deepgram](/leaderboard/claudecode/api/deepgram) | Voice STT | 86 | 2 |
| 4 | [AssemblyAI](/leaderboard/claudecode/api/assemblyai) | Voice STT | 86 | 2 |
| 5 | [Daytona](/leaderboard/claudecode/api/daytona) | Sandboxes | 86 | 2 |
| 6 | [Vercel](/leaderboard/claudecode/api/vercel) | Sandboxes | 85 | 2 |
| 7 | [OpenRouter](/leaderboard/claudecode/api/openrouter) | Inference | 85 | 2 |
| 8 | [Reducto](/leaderboard/claudecode/api/reducto) | Document Parsing | 84 | 2 |
| 9 | [Neo4j](/leaderboard/claudecode/api/neo4j) | Graph Databases | 83 | 2 |
| 10 | [Stripe](/leaderboard/claudecode/api/stripe) | Payment | 83 | 2 |

## Categories Sensitive To Agent Choice

Large gaps usually point to ambiguous docs, unusual auth flows, or task paths that one agent can recover from better than another. Categories with fewer than 6 rows are filtered out when possible.

| Category | Claude Code | Codex | Leader | Gap | Rows |
|----------|-------------|-------|--------|-----|------|
| Audio | 38 | 50 | Codex | 12 | 12 |
| E-Signature | 71 | 60 | Claude Code | 11 | 6 |
| Unified API | 59 | 52 | Claude Code | 7 | 8 |
| Voice Infra | 48 | 54 | Codex | 6 | 9 |
| Public Sector Intelligence | 25 | 30 | Codex | 5 | 8 |
| Stablecoin | 36 | 32 | Claude Code | 5 | 16 |
| Agent Automation | 38 | 34 | Claude Code | 4 | 6 |
| Voice Telephony | 43 | 47 | Codex | 4 | 19 |
| Video Infra | 68 | 71 | Codex | 3 | 6 |
| Agent Memory | 69 | 72 | Codex | 3 | 8 |
| Email | 71 | 74 | Codex | 3 | 6 |
| Payment | 51 | 48 | Claude Code | 2 | 24 |

## Category Market Leaders

API category leaders, closest challengers, and category lock-in.

| Category | Leader | Challenger | Gap | Companies |
|----------|--------|------------|-----|-----------|
| Agent Automation | [Dust](/leaderboard/claudecode/api/dust) (59) | Gumloop (25) | 34 | 3 |
| Audio | [Sonilo](/leaderboard/claudecode/api/sonilo) (81) | ElevenLabs Music (57) | 24 | 6 |
| Observability | [Datadog](/leaderboard/claudecode/api/datadog) (81) | BetterStack (65) | 17 | 3 |
| Graph Databases | [Neo4j](/leaderboard/claudecode/api/neo4j) (83) | Memgraph (69) | 14 | 4 |
| Meeting Bot | [Meetstream](/leaderboard/claudecode/api/meetstream) (78) | Meeting BaaS (65) | 13 | 5 |
| Voice TTS | [ElevenLabs](/leaderboard/claudecode/api/elevenlabs) (89) | Cartesia (78) | 11 | 6 |
| Public Sector Intelligence | [GovSpend](/leaderboard/claudecode/api/govspend) (44) | Pursuit (35) | 10 | 4 |
| CMS | [Shopware](/leaderboard/claudecode/api/shopware) (30) | Strapi (21) | 9 | 2 |
| Cloud Hosting | [Render](/leaderboard/claudecode/api/render) (77) | Cloudflare (69) | 8 | 7 |
| Payment | [Stripe](/leaderboard/claudecode/api/stripe) (83) | Mollie (76) | 7 | 12 |
| Document Parsing | [Reducto](/leaderboard/claudecode/api/reducto) (84) | Unstructured (78) | 6 | 5 |
| Durable Workflow | [Prefect](/leaderboard/claudecode/api/prefect) (78) | Restate (73) | 5 | 8 |
| Stablecoin | [DFNS](/leaderboard/claudecode/api/dfns) (63) | Coinbase Payments (58) | 5 | 8 |
| Inference | [OpenRouter](/leaderboard/claudecode/api/openrouter) (85) | Fireworks AI (81) | 4 | 14 |
| Webscraping | [fastCRW](/leaderboard/claudecode/api/fastcrw) (80) | Scrapfly (76) | 4 | 3 |

## Agent Effort And Cost

Average cost, runtime, tool calls, and errors on API company evals.

| Agent | Score | Avg Cost | Avg Runtime | Avg Tool Calls | Avg Errors | Rows |
|-------|-------|----------|-------------|----------------|------------|------|
| Claude Code | 55 | $0.49 | 4m 4s | 18 | 4 | 185 |
| Codex | 56 | $0.35 | 2m 13s | 23 | 4 | 182 |

## Friction Hotspots

API categories where agents spend the most effort recovering: more errors, longer runs, more tool calls, and higher cost.

| Category | Score | Avg Errors | Avg Runtime | Avg Cost | Avg Tool Calls | Rows |
|----------|-------|------------|-------------|----------|----------------|------|
| Agent Memory | 70 | 10 | 6m 50s | $0.78 | 43 | 8 |
| Video Infra | 70 | 6 | 4m 15s | $0.62 | 37 | 6 |
| Document Parsing | 73 | 6 | 2m 30s | $0.35 | 24 | 10 |
| Auth | 76 | 6 | 2m 17s | $0.44 | 30 | 12 |
| Stablecoin | 34 | 5 | 5m 18s | $0.80 | 24 | 16 |
| Email | 73 | 5 | 3m 9s | $0.56 | 30 | 6 |
| Cloud Hosting | 52 | 5 | 5m 9s | $0.75 | 30 | 14 |
| Browser | 77 | 5 | 2m 29s | $0.32 | 25 | 8 |
| Meeting Bot | 54 | 5 | 1m 49s | $0.34 | 18 | 10 |
| Verification | 49 | 5 | 1m 44s | $0.27 | 14 | 6 |
| Public Sector Intelligence | 27 | 5 | 3m 31s | $0.42 | 22 | 8 |
| Vector Databases | 67 | 4 | 3m 0s | $0.39 | 23 | 12 |

## Compare

Compare companies side by side at [/leaderboard/claudecode/api/compare](/leaderboard/claudecode/api/compare). Use `/leaderboard/{model}/{interface}/{company}` for individual provider details.

View the rest of the DevTool Arena results at [/leaderboard/claudecode/api](/leaderboard/claudecode/api).

## Frequently Asked Questions

### What does the discovery score measure?

The discovery score measures whether key agent-ready resources, such as llms.txt, Context7, MCP server documentation, typed SDKs, OpenAPI specs, agent skills, and CLI support, exist and are publicly available. Our methodology verifies the presence and accessibility of these resources; it is not a live test of whether an AI agent can independently find or use them in a task.

As a result, a company may score well on discovery because these assets are publicly available, even if an agent would not reliably find them in every real-world workflow.

### What does the eval score measure?

The eval score measures how well an AI coding agent can complete a real integration task using a company's API. We give the agent a practical prompt, such as generating a script, running it, and verifying the output, to see whether it can actually get the job done.

We then evaluate performance across the full workflow: finding the right docs, generating valid code, creating the necessary files, and executing the task successfully end to end.

We also capture supporting signals such as runtime, cost, tool calls, errors, and final execution outcome. The goal of the eval score is to measure real-world agent performance, not just theoretical API readiness.

### How is the final leaderboard score calculated?

The final leaderboard score combines the eval score and discovery score. The API leaderboard weights task execution at 70% and discovery at 30%, so agent-ready resources help the score, but successful end-to-end execution matters most.

