What each model was asked to do and how fast it did it: prefill, decode, time to first token, and the tail of the latency distribution.
Token mix
22.8B in · 106.2M out
22.87B total
Cache hit rate
0.0%
31.8B reused
Decode
-
weighted by output
Model
Requests
Tokens
Avg/req
Prefill
Decode
TTFT
Latency
Success
claude-opus-4-6
anthropic
61.2K
8.43B
8.4B20.9M
137.8K
-
-
-
-
avg -
100%
claude-opus-4-8
anthropic
58.2K
3.75B
3.7B35.2M
64.5K
-
-
-
-
avg -
100%
gpt-5.6-sol
openai
58
3.55B
3.5B14.2M
61.2M
-
-
-
-
avg -
100%
claude-opus-4-5-20251101
anthropic
2
3.23B
3.2B4.0M
1.6B
-
-
-
-
avg -
100%
claude-sonnet-4-6
anthropic
24.3K
877.82M
869.8M8.0M
36.1K
-
-
-
-
avg -
100%
claude-fable-5
anthropic
9.0K
809.67M
802.7M7.0M
89.8K
-
-
-
-
avg -
100%
gpt-5.5
openai
433
699.85M
696.0M3.1M
1.6M
-
-
-
-
avg 18530.16 s
100%
claude-haiku-4-5-20251001
anthropic
32.9K
638.28M
634.6M3.7M
19.4K
-
-
-
-
avg -
100%
claude-opus-5
anthropic
7.1K
435.95M
430.8M5.1M
61.2K
-
-
-
-
avg -
100%
claude-sonnet-5
anthropic
7.5K
305.39M
303.2M2.2M
40.8K
-
-
-
-
avg -
100%
grok-4.5
xai
129
35.74M
33.8M1.4M
277.0K
-
-
-
-
avg -
100%
claude-opus-4-7
anthropic
1.2K
33.92M
33.2M756.3K
27.2K
-
-
-
-
avg -
100%
gpt-5.6-luna
openai
6
29.51M
29.3M136.4K
4.9M
-
-
-
-
avg -
100%
gpt-5.4
openai
11
14.46M
14.4M64.9K
1.3M
-
-
-
-
avg -
100%
gpt-5.6-terra
openai
9
12.92M
12.8M108.2K
1.4M
-
-
-
-
avg -
100%
grok-4.5
openai
3
6.66M
6.3M231.7K
2.2M
-
-
-
-
avg -
100%
gpt-5.3-codex
openai
3
1.78M
1.8M14.0K
594.8K
-
-
-
-
avg -
100%
grok-composer-2.5-fast
xai
15
1.14M
1.1M42.5K
76.1K
-
-
-
-
avg -
100%
grok-4.5
auto
94
460.93K
383.0K77.9K
4.9K
-
-
-
-
avg -
100%
grok-composer-2.5-fast
openai
1
92.48K
90.4K1.8K
92.5K
-
-
-
-
avg -
100%
gpt-5.6
openai
2
76.85K
76.8K16
38.4K
-
-
-
-
avg -
100%
grok-composer-2.5
openai
2
76.85K
76.8K16
38.4K
-
-
-
-
avg -
100%
gpt-4o
openai
2
71.65K
71.6K10
35.8K
-
-
-
-
avg -
100%
gpt-4.1
openai
3
61.96K
61.9K21
20.7K
-
-
-
-
avg -
100%
gemini-2.5-pro
openai
2
49.34K
49.3K16
24.7K
-
-
-
-
avg -
100%
grok-build-0.1
xai
1
40.70K
40.3K201
40.7K
-
-
-
-
avg -
100%
grok-4.3
xai
1
40.49K
40.3K84
40.5K
-
-
-
-
avg -
100%
grok-4.20-0309-reasoning
xai
1
40.41K
40.1K133
40.4K
-
-
-
-
avg -
100%
grok-4.20-0309-non-reasoning
xai
1
40.15K
40.1K5
40.1K
-
-
-
-
avg -
100%
gpt-5
openai
2
39.46K
39.4K10
19.7K
-
-
-
-
avg -
100%
gpt-5.6-SOL
openai
1
38.42K
38.4K5
38.4K
-
-
-
-
avg -
100%
grok-4.20-multi-agent-0309
openai
1
38.42K
38.4K5
38.4K
-
-
-
-
avg -
100%
Luna
openai
2
37.96K
37.9K16
19.0K
-
-
-
-
avg -
100%
TERRA
openai
2
37.96K
37.9K16
19.0K
-
-
-
-
avg -
100%
claude-fable-5
openai
2
37.96K
37.9K16
19.0K
-
-
-
-
avg -
100%
claude-opus-4-8
openai
2
37.96K
37.9K16
19.0K
-
-
-
-
avg -
100%
claude-sonnet-5
openai
2
37.96K
37.9K16
19.0K
-
-
-
-
avg -
100%
o4-mini
openai
2
37.94K
37.9K10
19.0K
-
-
-
-
avg -
100%
claude-haiku-4-5-20251001
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-opus-4
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-opus-4-5-20251101
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-opus-4-6
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-opus-4-7
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-sonnet-4-5-20250929
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
claude-sonnet-4-6
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
o3-mini
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
sol
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
terra
openai
1
37.66K
37.7K5
37.7K
-
-
-
-
avg -
100%
gpt-5.6-Luna
openai
1
27.66K
27.7K5
27.7K
-
-
-
-
avg -
100%
gpt-5.6-TERRA
openai
1
27.66K
27.7K5
27.7K
-
-
-
-
avg -
100%
gemini-2.5-flash
openai
1
27.25K
27.2K5
27.3K
-
-
-
-
avg -
100%
SOL
openai
2
27.20K
27.2K16
13.6K
-
-
-
-
avg -
100%
claude-sonnet-4
openai
1
26.90K
26.9K5
26.9K
-
-
-
-
avg -
100%
claude-sonnet-4.5
openai
1
26.90K
26.9K5
26.9K
-
-
-
-
avg -
100%
o3
openai
1
26.90K
26.9K5
26.9K
-
-
-
-
avg -
100%
chatgpt-4o-latest
openai
1
25.07K
25.1K5
25.1K
-
-
-
-
avg -
100%
gpt-4.5
openai
1
25.07K
25.1K5
25.1K
-
-
-
-
avg -
100%
grok-composer-2.5-fast
auto
2
2.22K
1.7K552
1.1K
-
-
-
-
avg -
100%
gpt-5.6-instant
openai
1
1.04K
1.0K5
1.0K
-
-
-
-
avg -
100%
gpt-5.6-pro
openai
1
1.04K
1.0K5
1.0K
-
-
-
-
avg -
100%
gpt-5.6-thinking
openai
1
1.04K
1.0K5
1.0K
-
-
-
-
avg -
100%
gpt-4.5-preview
openai
1
497
4925
497
-
-
-
-
avg -
100%
gpt-5.5
auto
1
304
23569
304
-
-
-
-
avg -
100%
claude-opus-4-1-20250805
openai
1
280
2755
280
-
-
-
-
avg -
100%
luna
openai
1
280
2755
280
-
-
-
-
avg -
100%
Activity
When requests were sent, how large they were, and how long they took.
202.2K
Requests
0
Average tokens / request
-
Average TTFT
26195.22 s
Average latency
Requests by hour
00:0006:0012:0018:0023:00 UTC
Recent requests
No requests
Controller
Request volume grouped by the clients and providers passing through the controller.
By source
No controller data
By provider
No controller data
Errors
Failed requests reported by the controller during the past year.
0
Errors
100%
Success rate
202.2K
Requests observed
No errors in this period
About Tokens
This is my attempt to give people a 'state of my AI': a running account of which models I use and how much I use them, without requiring a separate explanation from me. Some session counts before June 2026 may be inaccurate. Although the lifetime totals have been reconciled from aggregate usage counters, the heatmap substantially underreports activity from December 25, 2025 through May 15, 2026. Claude Code was my primary tool during much of that period, but its retained daily token series does not begin until May 16, 2026.
