Terminus-2 – Kimi K3 (always on) vs. Claude Code – Claude Opus 5 (high): geospatial agent comparison

Comparison between Terminus-2 – Kimi K3 (always on) and Claude Code – Claude Opus 5 (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 – Kimi K3 (always on) vs. Claude Code – Claude Opus 5 (high)
Metric	Terminus-2 – Kimi K3 (always on)	Claude Code – Claude Opus 5 (high)
Geospatial Agent Index	58*	63* (partial: attempt 1 on 160 of 320 tasks)
Earth-Bench V2 score	61	63
GeoBenchX score	55	62
GeoAgentBench score (guided GIS workflow; outside the index)	94	100
GeoAnalystBench score (guided GIS workflow; outside the index)	89	100
Raster and remote sensing score	59	62
Vector and overlay score	33	No data
Networks and routing score	No data	No data
Climate and time series score	66	65
Spatial statistics and interpolation score	50	50
Mapping and cartography score	78	87
Recognising infeasible requests score	38	50
Cost per task	$0.336	$0.522
Time per task	3.9 min	2.3 min
Tokens per attempt	76k	253k
Turns per attempt	7.7	9.1
Attempts decided	320	160
Input price per 1M tokens	$3.00	$5.00
Output price per 1M tokens	$15.00	$25.00
Context window	1,000,000	No data
Image input	Yes	No data
Developer	Moonshot AI	Anthropic

Charts

Partial coverage. Claude Code – Claude Opus 5 (high): 112 of 147 Earth-Bench V2 tasks; 48 of 173 GeoBenchX tasks. Grey placeholders identify these entries in the index, cost and time 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

Planned attempts still to come

Data table: Geospatial Agent Index

Geospatial Agent Index
Model	Creator	Geospatial Agent Index	Range while attempts are pending	Benchmarks covered	Coverage (attempts)
Terminus-2 – Kimi K3 (always on)	Moonshot AI	58*	No pending judgements; coverage incomplete; 1 trace review declined by the judge	2 of 2	320 of 614 planned attempts

Cost per task

Average cost per task (USD) · Lower is better

Planned attempts still to come

How to read this chart

Not shown: partial coverage. Claude Code – Claude Opus 5 (high).

Data table: Cost per task

Cost per task
Model	Creator	Cost per task	Coverage (attempts)
Terminus-2 – Kimi K3 (always on)	Moonshot AI	$0.336	320 of 614 planned attempts

Time per task

Average agent wall time per task · Lower is better

Planned attempts still to come

How to read this chart

Not shown: partial coverage. Claude Code – Claude Opus 5 (high).

Data table: Time per task

Time per task
Model	Creator	Time per task	Coverage (attempts)
Terminus-2 – Kimi K3 (always on)	Moonshot AI	3.9 min	320 of 614 planned attempts

Frequently asked questions

Which scores higher on the Geospatial Agent Index, Terminus-2 – Kimi K3 (always on) or Claude Code – Claude Opus 5 (high)?

Not comparable yet: Claude Code – Claude Opus 5 (high) has partial coverage (112 of 147 Earth-Bench V2 tasks; 48 of 173 GeoBenchX tasks), so it is not ranked. Terminus-2 – Kimi K3 (always on): 320 of 614 planned attempts. Claude Code – Claude Opus 5 (high): 160 of 614 planned attempts. Coverage is partial; scores can change as more attempts are decided.

Which is cheaper per task, Terminus-2 – Kimi K3 (always on) or Claude Code – Claude Opus 5 (high)?

Terminus-2 – Kimi K3 (always on), at $0.336 against $0.522 for Claude Code – Claude Opus 5 (high).

Which is faster per task, Terminus-2 – Kimi K3 (always on) or Claude Code – Claude Opus 5 (high)?

Claude Code – Claude Opus 5 (high), at 2.3 min against 3.9 min for Terminus-2 – Kimi K3 (always on).

Which has the larger context window, Terminus-2 – Kimi K3 (always on) or Claude Code – Claude Opus 5 (high)?

Not known yet: Claude Code – Claude Opus 5 (high) has no context window data.

More: Terminus-2 – Kimi K3 (always on) · Claude Code – Claude Opus 5 (high) · all comparisons