Back

OpenAI

GPT-5.5 Instant (June 2026)

ReasoningClosed weights
Intelligence
26.0
Rank 164 of 744 measured models
Coding
39.4
Rank 156 of 298 measured models
Output speed
151 t/s
Rank 83 of 403 measured models
Price/1M
$11.25
Rank 630 of 659 measured models
Context window
400K
Rank 357 of 1199 measured models
Cost per task
$0.692
Rank 115 of 193 measured models

GPT-5.5 Instant (June 2026) scores 26.0 on the Artificial Analysis Intelligence Index, ranking 164th of 744 measured models. The closest models on this index are GLM-5.1 (26.1), MiMo-V2.5-Pro (26.0), and Kimi K2.7 Code (25.8). At $11.25 per million tokens (3:1 input/output blended price), it is more expensive than GLM-5.1 ($1.98), MiMo-V2.5-Pro ($0.544), and Kimi K2.7 Code ($1.71). With an output speed of 151 t/s, it is faster than GLM-5.1 (65.4 t/s), MiMo-V2.5-Pro (38.8 t/s), and Kimi K2.7 Code (58.2 t/s).

Artificial Analysis Indices
Intelligence26.0
Rank 164 of 744 measured models
Coding39.4
Rank 156 of 298 measured models
Agentic16.6
Rank 103 of 189 measured models
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting24.8
Rank 136 of 218 measured models
Strategy & Ops29.3
Rank 117 of 218 measured models
Legal29.5
Rank 122 of 218 measured models
Engineering24.7
Rank 138 of 217 measured models
Economics30.3
Rank 153 of 225 measured models
Standard Benchmarks
GPQA Diamond82.3%
Rank 182 of 619 measured models
HLE19.9%
Rank 219 of 664 measured models
SciCode52.5%
Rank 81 of 231 measured models
AA-LCR70.0%
Rank 218 of 575 measured models
Terminal-Bench 2.134.8%
Rank 155 of 244 measured models
Terminal-Bench v412.6%
Rank 77 of 218 measured models
τ-Bench Banking11.8%
Rank 141 of 211 measured models
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy