SOTA AI models ranked: LLM leaderboards and the kill line
The public AI leaderboards, collected daily: which models lead, and which cost more without scoring higher.
As of 2026-11-04: Gemini 4 Carbon leads Artificial Analysis Intelligence Index (60.0). Gemini 4 Carbon leads Arena — Text · Overall (style control) (1528). The kill line’s executioner is Claude Haiku 5.5, at a blended $0.20 per million tokens.
- SOTA
- 11models
- The executioner and every model above it
- No. 1 · AA
- Gemini 4 Carbon
- 60.0 · $8.00
- Executioner · price
- Claude Haiku 5.5
- $0.20
Ranking · AA 19–60 · ARENA 1417–1537Every model that outscores the executioner, with the executioner as the last row for reference. Each column is one board’s own score, shown side by side and never combined; a dash means that board does not list the model. Select a column head to sort by it. The rank is that board’s own; hover to see every board’s. A model marked Legacy is no longer listed as current by its source, and its row shows the best setting still listed.
- 160.01528
- 259.11509
- 356.91476
- 454.61481
- 554.51477
- 651.3—
- 750.41520
- 850.1—
- 950.01504
- 1046.5—
- 1143.61507
- 12Executioner36.51425
Kill line · score × price The executioner is the best value on the board: every model that costs more and scores no higher is killed.Each dot is a model: further right costs more, higher scores more. The dashed lines cross at the executioner. Above the horizontal line is SOTA, the top tier; below and to the left is cheaper and weaker, the low-cost tier.24 of 38 models are killed. The faint line joins the models that no other model beats on both price and score. Legacy models are not plotted. Select a model to see its details below.How the kill line is drawn →Who held it before →
The executioner is the best value on the board: every model that costs more and scores no higher is killed. How the kill line is drawn →Who held it before →
PricePer 1M tokens: the list price, independent of effort. Per task: the average cost of running one Intelligence Index task, which rises with effort, so a model’s effort settings spread horizontally. Models without a published cost per task are omitted.
Kill lineFrontier: models that no other model beats on both price and scoreSOTALow-costKilledHollow dot or faint cross — estimated by the sourceShort dashes — the selected model at its other effort settings
Gemini 4 Carbon
Scores on each board Each rank is the board’s own: the rank it publishes, or the model’s place in the board’s own order where it publishes none. The bar shows how much of that board the model ranks ahead of. Scores from different boards cannot be compared.
- Artificial Analysis LLM Leaderboard60.0No. 1 of 43
- Arena — Agent0.166No. 2 of 57
- Artificial Analysis Coding Agents0.594No. 4 of 27
- Arena — Code1768No. 4 of 152
- Arena — Text1528No. 1 of 419
Scores at each effort setting The same model at each effort setting. Each bar runs from zero to that board’s best score, so read each board separately, never across boards.
Artificial Analysis LLM Leaderboardbest at high
- low54.2
- medium58.3
- high60.0
Tested at one setting only
- Arena — Agent · unknownhigh0.166
- Artificial Analysis Coding Agents · Antigravity CLIhigh0.594
- Arena — Codehigh1768
- Arena — Texthigh1528