GPT-5.5
openai/gpt-5.5
152.3
median tok/s
Throughput Runs
8
TTFT Runs
6
Avg TTFT
1072ms
Avg Throughput
193.3 tok/s
Total Cost
$0.9330
Commentary
by derivedGPT-5.5 posts 152.3 tok/s median throughput with 930ms median TTFT across 15 runs. Success rate is 100.0% with $0.93 total spend.
Startup latency on ttft-factual landed in the 930ms range.
Startup latency on ttft-definition landed in the 930ms range.
Sustained decode speed on throughput-data-structures contributed to the 152.3 tok/s median.
Sustained decode speed on throughput-api-design contributed to the 152.3 tok/s median.
Sustained decode speed on throughput-essay contributed to the 152.3 tok/s median.
Notable Prompts
Fastest throughput run peaked at 468.7 tok/s.
Slowest startup path took 1786ms to first token.
All Runs
| Prompt | Type | Tok/s | TTFT | Tokens | Cost | |
|---|---|---|---|---|---|---|
1. Api Design throughput-api-design | throughput | 468.7 | 44864ms | 4096 | $0.1234 | |
2. Api Design throughput-api-design | throughput | 104.7 | 9761ms | 4096 | $0.1234 | |
3. Api Design throughput-api-design | throughput | 302.4 | 31928ms | 4096 | $0.1234 | |
1. Data Structures throughput-data-structures | throughput | 185.8 | 14849ms | 3720 | $0.1122 | |
2. Data Structures throughput-data-structures | throughput | 146.9 | 6050ms | 3002 | $0.0906 | |
3. Data Structures throughput-data-structures | throughput | 157.8 | 8833ms | 3347 | $0.1010 | |
1. Essay throughput-essay | throughput | 98.2 | 15355ms | 4096 | $0.1234 | |
2. Essay throughput-essay | throughput | n/a | 4735ms | 199 | $0.0060 | |
3. Essay throughput-essay | throughput | 81.8 | 3503ms | 4096 | $0.1234 | |
1. Definition ttft-definition | ttft | n/a | 986ms | 51 | $0.0016 | |
2. Definition ttft-definition | ttft | n/a | 1786ms | 48 | $0.0015 | |
3. Definition ttft-definition | ttft | n/a | 890ms | 53 | $0.0017 | |
1. Factual ttft-factual | ttft | n/a | 949ms | 14 | $0.0005 | |
2. Factual ttft-factual | ttft | n/a | 909ms | 14 | $0.0005 | |
3. Factual ttft-factual | ttft | n/a | 910ms | 14 | $0.0005 |
15 runs · Throughput rows require valid long-output runs · TTFT shown for all successful runs