# Company frontier share

Replayed using 2026-10-04 scores and prices, not historical measurements.

| Rank | Company | Share | Models | Rank change |
| --- | --- | --- | --- | --- |
| 1 | OpenAI | 61.70% | 2 | — |
| 2 | Anthropic | 29.44% | 1 | — |
| 3 | Xiaomi | 8.86% | 2 | — |
Budget: $10.63–$20000 per benchmark run.


## Overall leaderboard

Average share across this range

2025-10-04–2026-10-04

| Rank | Company | Average share |
| --- | --- | --- |
| 1 | OpenAI | 32.08% |
| 2 | Anthropic | 27.89% |
| 3 | DeepSeek | 20.92% |
| 4 | Mistral | 9.01% |
| 5 | Google | 4.38% |
| 6 | Xiaomi | 2.90% |
| 7 | Z AI | 1.73% |
| 8 | Alibaba | 0.37% |
| 9 | IBM | 0.26% |
| 10 | Meta | 0.23% |
| 11 | SpaceXAI | 0.18% |
| 12 | StepFun | 0.06% |

## How we compare models

Intelligence uses the Artificial Analysis Intelligence Index. Benchmark cost is the total USD cost of running that test, including reasoning tokens; it is not a per-token API price. Rankings use the same dataset and calculations in every language. General benchmark scores do not measure performance in a particular language or task.

Data from Artificial Analysis and other sources. Independent project; not affiliated with or endorsed by Artificial Analysis.


> As known by 2026-10-04T15:28:36.039005192Z. Last observation: 2026-10-04T15:28:36.039005192Z (refresh). Values carry forward from that observation; missing days are not new measurements. Metrics snapshot: 2026-10-04T15:28:36Z. Intelligence index version: unknown. Release timeline, frontier age, release impact and shakeup cards are reconstructions using one selected snapshot's metrics, not a record of what was known at release. Scores, benchmark cost and derived value may change when the index methodology changes. Source documentation advertises v4.3.2; API correspondence is unverified.