Terminus-2 – gpt-oss-20b (high) vs. Codex – GPT-6 Luna (high): geospatial agent comparison

Comparison between Terminus-2 – gpt-oss-20b (high) and Codex – GPT-6 Luna (high) on the Geospatial Agent Index, each benchmark, cost, time and token use, from the same tasks, agent and limits. Methodology.

Side by side

Terminus-2 – gpt-oss-20b (high) vs. Codex – GPT-6 Luna (high)
Metric	Terminus-2 – gpt-oss-20b (high)	Codex – GPT-6 Luna (high)
Geospatial Agent Index	34* (partial: attempt 1 on 477 of 488 tasks)	71* (partial: attempt 1 on 444 of 488 tasks)
Earth-Bench V2 score	32	66
GeoBenchX score	38	71
GS-QA2 score	30	77
GeoAgentBench score (guided GIS workflow; outside the index)	53	96
GeoAnalystBench score (guided GIS workflow; outside the index)	37	100
Raster and remote sensing score	34	73
Vector and overlay score	27	77
Climate and time series score	31	69
Spatial statistics and interpolation score	0	33
Mapping and cartography score	21	71
Recognising infeasible requests score	67	78
Cost per task	$0.017	$0.0039
Time per task	11.8 min	1.7 min
Tokens per attempt	67k	116k
Turns per attempt	5.2	7.3
Attempts decided	477	430
Input price per 1M tokens	$0.20	$0.10
Output price per 1M tokens	$0.30	$0.50
Context window	128,000	No data
Image input	No	No data
Developer	OpenAI	OpenAI

Charts

Partial coverage. Terminus-2 – gpt-oss-20b (high): 210 of 221 GS-QA2 tasks; Codex – GPT-6 Luna (high): 106 of 150 GeoBenchX tasks. Grey placeholders identify these entries in the index, cost and output token charts; their height does not represent a value. Comparable bars appear when first-attempt coverage is complete. Their available results remain in the side-by-side table.

Geospatial Agent Index

Equal-weight average of the evaluation scores, 0 to 100 · Higher is better

Data table: Geospatial Agent Index

Geospatial Agent Index
Model	Creator	Geospatial Agent Index	Range while attempts are pending	Benchmarks covered	Coverage (attempts)

Cost per task

Average cost per task (USD) · Lower is better

How to read this chart

Not shown: partial coverage. 2 entries without a first attempt on every index task. What is still running is on the models page.

Data table: Cost per task

Cost per task
Model	Creator	Cost per task	Coverage (attempts)

Output tokens per task

Average output tokens per attempt, reasoning included · Lower is better

How to read this chart

Not shown: partial coverage. 2 entries without a first attempt on every index task. What is still running is on the models page.

Data table: Output tokens per task

Output tokens per task
Model	Creator	Output tokens per task	Coverage (attempts)

Frequently asked questions

Which scores higher on the Geospatial Agent Index, Terminus-2 – gpt-oss-20b (high) or Codex – GPT-6 Luna (high)?

Not comparable yet: Terminus-2 – gpt-oss-20b (high) has partial coverage (210 of 221 GS-QA2 tasks) and Codex – GPT-6 Luna (high) has partial coverage (106 of 150 GeoBenchX tasks), so it is not ranked. Terminus-2 – gpt-oss-20b (high): 477 of 722 scheduled attempts (6 excluded). Codex – GPT-6 Luna (high): 430 of 722 scheduled attempts. Coverage is partial; scores can change as more attempts are decided.

Which is cheaper per task, Terminus-2 – gpt-oss-20b (high) or Codex – GPT-6 Luna (high)?

Codex – GPT-6 Luna (high), at $0.0039 against $0.017 for Terminus-2 – gpt-oss-20b (high).

Which writes fewer output tokens per task, Terminus-2 – gpt-oss-20b (high) or Codex – GPT-6 Luna (high)?

Codex – GPT-6 Luna (high), at 3k against 40k for Terminus-2 – gpt-oss-20b (high).

Which has the larger context window, Terminus-2 – gpt-oss-20b (high) or Codex – GPT-6 Luna (high)?

Not known yet: Codex – GPT-6 Luna (high) has no context window data.

More: Terminus-2 – gpt-oss-20b (high) · Codex – GPT-6 Luna (high) · all comparisons