FrontierMax · default fallback
For complex coding, careful analysis, and work that needs sustained judgment.
AA intelligence58
Speed · tok/s89.8
Output / 1M$20
Context1M
A CLEARER VIEW OF SI
Find the right model.
Understand the tradeoffs. Compare the benchmarks. Build with confidence.
12 current models8 leading labsVerified Oct 1, 2026
The model landscape
Capability is only part of the picture.
⌘ K
| Compare | # | AA index · v4.3.2 | tokens / sec | USD / 1M tokens | USD / 1M tokens | tokens | Model details | |
|---|---|---|---|---|---|---|---|---|
| 1 | 58 | 89.8 | $4 | $20 | 1M | |||
| 2 | 56 | 139.0 | $2 | $10 | 1M | |||
| 3 | 53 | — | $2 | $10intro | 1M | |||
| 3 | 53 | 51.3 | $10 | $50 | 1.05M | |||
| 5 | 52 | 64.0 | $2 | $10 | 1.05M | |||
| 6 | 48 | 157.1 | $1.25 | $4.25 | 1.05M | |||
| 7 | 46 | 74.6 | $2 | $6 | 500K | |||
| 8 | 41 | 223.8 | $0.75 | $3.75 | 1M | |||
| 9 | 39 | 209.2 | $0.3 | $1.2 | 1M | |||
| 10 | 38 | 125.2 | $0.1 | $0.5 | 1.05M | |||
| 11 | 14 | 164.7 | $1.5 | $7.5 | 256K | |||
| — | — | — | — | — | — |
NewMax · default fallback
A fast iteration partner for everyday coding, documents, and visual work.
AA intelligence56
Speed · tok/s139.0
Output / 1M$10
Context1M
Early accessHigh
A new frontier model for long workflows, visual analysis, and cyber defense.
AA intelligence53
Speed · tok/s—
Output / 1M$10
Context1M
FrontierMax
For demanding reasoning, scientific research, and complex agent workflows.
AA intelligence53
Speed · tok/s51.3
Output / 1M$50
Context1.05M
NewMax
A balanced starting point for capable coding, computer use, and professional work.
AA intelligence52
Speed · tok/s64.0
Output / 1M$10
Context1.05M
Max
For multi-step agents and workflows that combine documents, images, and video.
AA intelligence48
Speed · tok/s157.1
Output / 1M$4.25
Context1.05M
Xhigh
For coding and knowledge work with a lower output price than several frontier peers.
AA intelligence46
Speed · tok/s74.6
Output / 1M$6
Context500K
High
For responsive applications and inputs spanning text, images, audio, and video.
AA intelligence41
Speed · tok/s223.8
Output / 1M$3.75
Context1M
Open weightsMax
For economical API workloads or self-hosted agents with open model weights.
AA intelligence39
Speed · tok/s209.2
Output / 1M$1.2
Context1M
Max
For focused, high-volume tasks where the API budget is a primary constraint.
AA intelligence38
Speed · tok/s125.2
Output / 1M$0.5
Context1.05M
AA published setting
For customizable coding and agent applications with deployable model weights.
AA intelligence14
Speed · tok/s164.7
Output / 1M$7.5
Context256K
UnrankedNot verified
An omnimodal model announced for tool-using agents and creative workflows.
AA intelligence—
Speed · tok/s—
Output / 1M—
Context—
Independent scores from Artificial Analysis. Prices are standard text API rates.
Capability meets cost
A better model for your budget starts here.
Open weightsProprietaryHigher & further left = stronger value
BEYOND THE LEADERBOARD
Great benchmarks.
Better decisions.
The right model depends on your work. A faster, less expensive model can be the better choice for a focused task.
Compare the tradeoffsCost, context, access, and capability.
Test on your own tasksBenchmarks guide. Your workload decides.