THE FORGE RANKINGS·2026-08-30

GPT-5.6 Terra

OpenAI
12
RANK of 37 ranked
FERROX INDEX80.1 weighted mean of category positions
GRADEA front rank
EVIDENCE8 / 32 benchmarks measured
QUALITYFULL every category rests on independent benchmarks

Not separable from GPT-5.5 and Grok 4.5 at the measured 1.78 point band. The order is printed; the gap is not claimed.

No measurement in Preference. That weight was redistributed, which is the same as assuming this model would have scored its own average there. It is an assumption, not a measurement.

Compared to what

Each rail is one classification. Every ranked model measured on it is a tick; this model is the orange marker. A hatched rail means this model has no measurement in that classification and its weight was redistributed.

Agents

34of 37 measured
24.1
13.1median 38.963.0

Coding

18of 37 measured
50.3
22.4median 50.070.7

Reasoning

9of 37 measured
81.1
18.1median 62.893.0

Preference

no qualifying measurement
53.1median 71.882.5

Every measurement, and the field behind it

One card per benchmark this model has been measured on. The rail shows where it sits against every other ranked model measured on the same benchmark. Where the evaluator keys on a harness, the harness is named, because the same model scores differently under a different one.

ALE-Bench

4/35
1951 best 2177 · GPT-5.6 Sol
ALE-Bench (Sakana AI with AtCoder) CC-BY-4.0 max

LMArena Agent

14/21
0.030 best 0.127 · Claude Opus 5
LMArena CC-BY-4.0

FrontierCode

10/17
41.3% best 53.5% · Claude Fable 5
cognition.com CC-BY-4.0 codex max

Terminal-Bench 4.0

6/10
21.5% best 51.8% · Claude Opus 5
Terminal-Bench Apache-2.0 codex max

Not measured

24/32
  • APEX Agents
  • FrontierMath
  • SWE-bench Verified (Epoch's own run)
  • LMArena Text (style-controlled)
  • Humanity's Last Exam
  • Terminal-Bench 2
  • LMArena WebDev
  • METR time horizons
  • Aider Polyglot
  • OSWorld
  • Cybench
  • BrowseComp
  • IMOAnswerBench
  • SWE-bench Multilingual
  • AIME
  • CyberGym
  • HMMT
  • MCP-Atlas
  • SWE-bench Pro
  • Tau2-Bench Airline
  • Tau2-Bench Banking
  • Tau2-Bench Retail
  • Tau2-Bench Telecom
  • Tool-Decathlon
A missing benchmark is not a zero and is never scored as one.

Check it yourself

Every number on this page, with who measured it, the interval they published, the harness it was run under, and where to go and read it. Nothing here was measured by Ferrox Labs.

BenchmarkClassPublishedInterval (normalised) Measured byLicenceHarness / effort
LMArena Agent agents 0.030 22.7 to 30.7 LMArena CC-BY-4.0 source
Terminal-Bench 4.0 agents 21.5% 18.3 to 24.8 Terminal-Bench Apache-2.0 codex / max source
ALE-Bench coding 1951 not published ALE-Bench (Sakana AI with AtCoder) CC-BY-4.0 / max source
DeepSWE coding 53.8% 49.4 to 58.1 deepswe.datacurve.ai CC-BY-4.0 mini-swe-agent / high source
FrontierCode coding 41.3% not published cognition.com CC-BY-4.0 codex / max source
ARC-AGI-2 reasoning 67.1% not published ARC Prize CC-BY-4.0 / high source
GPQA Diamond (Epoch's own run) reasoning 87.4% 82.7 to 92.0 Epoch AI CC-BY-4.0 / low source
OTIS Mock AIME 2024-2025 reasoning 88.9% 79.6 to 98.2 Epoch AI CC-BY-4.0 / low source

Configurations rolled up: 12. Rule: median observed configuration per benchmark (lower median, always a real measurement). Harnesses seen: codex, mini-swe-agent.

Snapshot 2026-08-30-62e4e85f043b, manifest dc46131315046f1e. Grades are positional: position in the measured field, as a percentile of rank among ranked entries, n=37.