Terminus-2 – GLM-5.3 (max, default) vs. Claude Code – Claude Haiku 4.5 (thinking 32k): geospatial agent comparison

Comparison between Terminus-2 – GLM-5.3 (max, default) and Claude Code – Claude Haiku 4.5 (thinking 32k) 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 – GLM-5.3 (max, default) vs. Claude Code – Claude Haiku 4.5 (thinking 32k)
Metric	Terminus-2 – GLM-5.3 (max, default)	Claude Code – Claude Haiku 4.5 (thinking 32k)
Geospatial Agent Index	38*	36* (partial: attempt 1 on 94 of 320 tasks)
Earth-Bench V2 score	33	28
GeoBenchX score	42	45
GeoAgentBench score (guided GIS workflow; outside the index)	92	69
GeoAnalystBench score (guided GIS workflow; outside the index)	84	100
Raster and remote sensing score	31	33
Vector and overlay score	50	No data
Networks and routing score	No data	No data
Climate and time series score	38	33
Spatial statistics and interpolation score	0	40
Mapping and cartography score	58	60
Recognising infeasible requests score	38	26
Cost per task	$0.525	$0.155
Time per task	12.4 min	2.5 min
Tokens per attempt	385k	490k
Turns per attempt	12.2	14.2
Attempts decided	320	94
Input price per 1M tokens	$1.40	$1.00
Output price per 1M tokens	$4.40	$5.00
Context window	1,048,576	200,000
Image input	No	Yes
Developer	Z.ai	Anthropic

Charts

Partial coverage. Claude Code – Claude Haiku 4.5 (thinking 32k): 47 of 147 Earth-Bench V2 tasks; 47 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 – GLM-5.3 (max, default)	Z.ai	38*	No pending judgements; coverage incomplete	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 Haiku 4.5 (thinking 32k).

Data table: Cost per task

Cost per task
Model	Creator	Cost per task	Coverage (attempts)
Terminus-2 – GLM-5.3 (max, default)	Z.ai	$0.525	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 Haiku 4.5 (thinking 32k).

Data table: Time per task

Time per task
Model	Creator	Time per task	Coverage (attempts)
Terminus-2 – GLM-5.3 (max, default)	Z.ai	12.4 min	320 of 614 planned attempts

Frequently asked questions

Which scores higher on the Geospatial Agent Index, Terminus-2 – GLM-5.3 (max, default) or Claude Code – Claude Haiku 4.5 (thinking 32k)?

Not comparable yet: Claude Code – Claude Haiku 4.5 (thinking 32k) has partial coverage (47 of 147 Earth-Bench V2 tasks; 47 of 173 GeoBenchX tasks), so it is not ranked. Terminus-2 – GLM-5.3 (max, default): 320 of 614 planned attempts. Claude Code – Claude Haiku 4.5 (thinking 32k): 94 of 614 planned attempts. Coverage is partial; scores can change as more attempts are decided.

Which is cheaper per task, Terminus-2 – GLM-5.3 (max, default) or Claude Code – Claude Haiku 4.5 (thinking 32k)?

Claude Code – Claude Haiku 4.5 (thinking 32k), at $0.155 against $0.525 for Terminus-2 – GLM-5.3 (max, default).

Which is faster per task, Terminus-2 – GLM-5.3 (max, default) or Claude Code – Claude Haiku 4.5 (thinking 32k)?

Claude Code – Claude Haiku 4.5 (thinking 32k), at 2.5 min against 12.4 min for Terminus-2 – GLM-5.3 (max, default).

Which has the larger context window, Terminus-2 – GLM-5.3 (max, default) or Claude Code – Claude Haiku 4.5 (thinking 32k)?

Terminus-2 – GLM-5.3 (max, default), at 1,048,576 tokens against 200,000 tokens for Claude Code – Claude Haiku 4.5 (thinking 32k).

More: Terminus-2 – GLM-5.3 (max, default) · Claude Code – Claude Haiku 4.5 (thinking 32k) · all comparisons