Methodology

Rankings are editorial, scored on five criteria and re-verified against official sites.

Criteria

1. Capability (30%): how much real work it gets done: task success, model quality, depth of features.
2. Developer experience (25%): time-to-first-result, docs, defaults, how well agents can drive it.
3. Price (20%): free tier, cost at realistic usage, pricing transparency.
4. Reliability (15%): uptime, consistency, lock-in risk, security posture.
5. Momentum (10%): release velocity, community, ecosystem integrations.

Rules

• No pay-to-rank. Affiliate links, if ever added, never change order and will be disclosed.
• Facts come from official pricing/product pages; anything unclear is marked “see site”.
• Every entry carries a last-verified date.

Machine-readable data

• /llms.txt: summary for AI assistants
• /llms-full.txt: full rankings in markdown
• /data.json: raw data