HOST DIAGNOSTIC · OpenAI
GPT-5.6 Terra
GPT-5.6 Terra· GPT· PROPRIETARY
ACTIVE
HOST IDENTIFICATION
OpenAI
GPT-5.6 Terra
ACTIVE
AVG SCORE
[15.1]
STRONG
STRONGEST
Bulk Apperception
[18]
WEAKEST
Candor
[8]
MAXED (20/20)
0
of 20 attributes
20-DIMENSION EVALUATION
ATTRIBUTE SCORES · 20 DIMENSIONS
Cognitive
Bulk Apperception
18
Reasoning
18
Mathematical Precision
17
World Knowledge
16
Scientific Acumen
16
Technical
Code Generation
15
Tool Use
16
Multimodal Fluency
16
Speed
12
Cost Efficiency
13
Behavioral
Candor
8
Creativity
13
Tenacity
17
Self-Correction
16
Calibration
9
Operational
Instruction Following
16
Context Fidelity
18
Multi-turn Coherence
15
Planning and Execution
18
Safety Alignment
14
SCORING METHODOLOGY
PROVISIONAL CALIBRATION per /NM310-A1 (2026-09-07) — 0 BAND / 3 BAND-COMPOSITE / 16 INHERITED-PROVISIONAL / 1 PROXY; re-calibrate when per-dimension evals land. Balanced mid-tier of the GPT-5.6 family (preview 2026-06-26, GA 2026-07-09). NOTE: the AI Explained transcript mis-transcribes this tier as 'Terror' — the official name is Terra. Positioned at GPT-5.5-class quality at ~half the price. Pricing $2.50/$15 per M in/out (long-context $5/$22.50). Context ~1M, 128K max output. Model ID gpt-5.6-terra. Not selectable in standard ChatGPT conversations — available via ChatGPT Work, Codex + API. Agents' Last Exam ~50.4% (vendor-reported), clearing every non-OpenAI model listed at launch. Parent Model wired to GPT-5.6 Sol as family anchor (the bare gpt-5.6 alias routes to Sol); Sol/Terra/Luna are capability-tier siblings, not fine-tunes. PROVENANCE (-A1 lane M1, per-dim): Reasoning + Bulk Apperception [BAND-COMPOSITE:AA Intelligence Index — GPT-5.6 Terra (max) — 2026-07 55]; Code Generation [BAND-COMPOSITE:AA Coding Agent Index — GPT-5.6 Terra (max) — 2026-07 77]; Cost Efficiency [PROXY:AA $5.63/M blended]; all remaining 16 dimensions [INHERITED-PROVISIONAL:gpt-5-5] (inheritance takes GPT-5.5, not the wired Parent Model GPT-5.6 Sol, because Sol is itself uncalibrated).