Back

OpenAI

GPT-5.6 Terra

ReasoningClosed weights
Intelligence
34.2
Rank 98 of 744 measured models
Coding
67.1
Rank 64 of 298 measured models
Output speed
92.9 t/s
Rank 191 of 403 measured models
Price/1M
$4.50
Rank 563 of 659 measured models
Context window
1M
Rank 84 of 1199 measured models
Cost per task
$0.338
Rank 90 of 193 measured models

GPT-5.6 Terra (High) scores 34.2 on the Artificial Analysis Intelligence Index, ranking 98th of 744 measured models. The closest models on this index are DeepSeek V4 Flash 0731 (Max) (34.3), Gemini 3.6 Flash (High) (34.0), and JT-4.1 Flash 236B A21B (33.9). At $4.50 per million tokens (3:1 input/output blended price), it is more expensive than DeepSeek V4 Flash 0731 (Max) ($0.66) and Gemini 3.6 Flash (High) ($1.50). With an output speed of 92.9 t/s, it is slower than DeepSeek V4 Flash 0731 (Max) (173 t/s) and Gemini 3.6 Flash (High) (175 t/s).

Artificial Analysis Indices
High
Intelligence34.2
Rank 98 of 744 measured models
Coding67.1
Rank 64 of 298 measured models
Agentic36.7
Rank 55 of 189 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 & Accounting37.1
Rank 95 of 218 measured models
Strategy & Ops40.0
Rank 89 of 218 measured models
Legal37.0
Rank 93 of 218 measured models
Healthcare & Medical34.1
Rank 47 of 82 measured models
Engineering34.3
Rank 96 of 217 measured models
Economics41.6
Rank 101 of 225 measured models
Standard Benchmarks
GPQA Diamond89.6%
Rank 74 of 619 measured models
HLE38.5%
Rank 96 of 664 measured models
SciCode52.4%
Rank 82 of 231 measured models
IFBench64.4%
Rank 109 of 450 measured models
AA-LCR77.7%
Rank 126 of 575 measured models
TerminalBench Hard57.6%
Rank 11 of 432 measured models
Terminal-Bench 2.175.7%
Rank 62 of 244 measured models
Terminal-Bench v41.5%
Rank 118 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