State of the market

Trends

Aggregate view of the 385-model catalog: what capabilities are standard, where prices sit, which labs charge what, and what has been moving lately. Computed at build time from the same open snapshot as every other page.

Canonical market: 385; tracked models
Median rate · chat: $0.473; per 1M billable tokens
Median context: 262K; known windows
Tape · last 24 days: 957; catalog changes

Capability adoption

reasoning
254 · 66%
tool call
332 · 86%
structured output
280 · 73%
vision / attachments
227 · 59%
audio input
61 · 16%
open weights
186 · 48%

Price distribution

chat rate /M · 351 priced models

$0
18
< $0.25
123
$0.25 – $1
105
$1 – $3
57
$3 – $10
45
$10 – $30
14
≥ $30
7

Median prices by lab

Labs with ≥ 3 priced models. Log-scaled bars keep lower-priced labs readable; the printed prices remain authoritative.

Input medianOutput median

openai
in $1.25
out $8.50
59 mod.
alibaba
in $0.14
out $0.58
49 mod.
google
in $0.38
out $1.88
43 mod.
mistral
in $0.20
out $0.57
31 mod.
anthropic
in $2.70
out $13.50
22 mod.
deepseek
in $0.18
out $0.33
17 mod.
zhipuai
in $0.29
out $0.96
16 mod.
meta
in $0.14
out $0.40
13 mod.

Context vs price explorer

Find the models that buy the most context for the least published price. Switch between input, output and the site’s 75/25 blend; then slice by lab, weight access, price or context. Both axes are logarithmic.

344 visible · 5 Pareto-efficient

1×

Hosted weightsOpen weightsPareto frontierBands: budget ≤ $0.25 · mainstream $0.25–$3 · premium ≥ $3

Select or focus a dot to pin its full input, output, context and provider details.

Change pulse

repriced
1142
new model
892
capability
725
context
682
removed
456
new provider
127
deprecated
14