Back

OpenAI

GPT-6 Sol

Closed weights
Intelligence
28.5
Rank 137 of 744 measured models
Output speed
86.4 t/s
Rank 206 of 403 measured models
Price/1M
$4
Rank 532 of 659 measured models
Context window
1.1M
Rank 21 of 1199 measured models
Cost per task
$0.334
Rank 89 of 193 measured models

GPT-6 Sol (Non-reasoning) scores 28.5 on the Artificial Analysis Intelligence Index, ranking 137th of 744 measured models. The closest models on this index are MiMo-V2-Pro (28.6), GPT-5.2 Codex (Xhigh) (28.5), and Qwen3.6 Max Preview (28.4). At $4 per million tokens (3:1 input/output blended price), it is cheaper than GPT-5.2 Codex (Xhigh) ($4.81), more expensive than Qwen3.6 Max Preview ($2.93). With an output speed of 86.4 t/s, it is faster than Qwen3.6 Max Preview (62.3 t/s), slower than GPT-5.2 Codex (Xhigh) (115 t/s).

Artificial Analysis Indices
Non-reasoning
Intelligence28.5
Rank 137 of 744 measured models
Reasoning levels
Each point is a reasoning level of this model: more reasoning often raises the score but costs more. Select a point to open that level.
Intelligence Indexmost efficient ↗Cost per task
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting32.0
Rank 108 of 218 measured models
Strategy & Ops35.8
Rank 103 of 218 measured models
Legal32.4
Rank 110 of 218 measured models
Engineering29.1
Rank 117 of 217 measured models
Economics34.7
Rank 134 of 225 measured models
Standard Benchmarks
HLE18.4%
Rank 236 of 664 measured models
SciCode47.3%
Rank 129 of 231 measured models
AA-LCR64.0%
Rank 264 of 575 measured models
Terminal-Bench v413.1%
Rank 75 of 218 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