Back

OpenAI

GPT-6 Sol

ReasoningClosed weightsModels.dev
Intelligence
47.6
Rank 28 of 744 measured models
Output speed
94.1 t/s
Rank 183 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
$1.04
Rank 133 of 193 measured models

GPT-6 Sol (Max) scores 47.6 on the Artificial Analysis Intelligence Index, ranking 28th of 744 measured models. The closest models on this index are Muse Spark 1.3 (48.1), GPT-5.6 Sol (47.0), and Grok 4.7 (46.4). At $4 per million tokens (3:1 input/output blended price), it is cheaper than GPT-5.6 Sol ($8), more expensive than Muse Spark 1.3 ($2) and Grok 4.7 ($3). With an output speed of 94.1 t/s, it is faster than GPT-5.6 Sol (85.8 t/s) and Grok 4.7 (78.7 t/s), slower than Muse Spark 1.3 (129 t/s).

Artificial Analysis Indices
Maximum
Intelligence47.6
Rank 28 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 & Accounting48.6
Rank 34 of 218 measured models
Strategy & Ops52.5
Rank 29 of 218 measured models
Legal51.1
Rank 35 of 218 measured models
Healthcare & Medical43.6
Rank 22 of 82 measured models
Engineering48.6
Rank 29 of 217 measured models
Economics54.4
Rank 34 of 225 measured models
Cyber Index
Artificial Analysis · v1
Cyber Index36.8%Rank 11 of 20 measured models
Successes
36.8%
Safety blocks
36.4%
Failures
26.8%
CWE-Bench-AA64.2% (safety blocks)
DeepsecBench-AA46.3%
CyberGym-E2E-AA0.0% (safety blocks)
Cost per task$1.96

* The model or provider declined some tasks on safety grounds.

Standard Benchmarks
HLE47.9%
Rank 34 of 664 measured models
SciCode57.6%
Rank 25 of 231 measured models
AA-LCR83.7%
Rank 15 of 575 measured models
Terminal-Bench v443.9%
Rank 25 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