Terminus-2 – DeepSeek V4 Pro (high) vs. Claude Code – GLM-5.3 (max, default): geospatial agent comparison

Comparison between Terminus-2 – DeepSeek V4 Pro (high) and Claude Code – GLM-5.3 (max, default) 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 – DeepSeek V4 Pro (high) vs. Claude Code – GLM-5.3 (max, default)
Metric	Terminus-2 – DeepSeek V4 Pro (high)	Claude Code – GLM-5.3 (max, default)
Geospatial Agent Index	63* (partial: attempt 1 on 559 of 565 tasks)	64* (partial: attempt 1 on 117 of 565 tasks)
Earth-Bench V2 score	61	76
GeoBenchX score	57	52
GS-QA2 score	71	No data
GeoAgentBench score (guided GIS workflow; outside the index)	96	No data
GeoAnalystBench score (guided GIS workflow; outside the index)	100	No data
Raster and remote sensing score	71	68
Vector and overlay score	64	67
Networks and routing score	No data	No data
Climate and time series score	62	100
Spatial statistics and interpolation score	21	33
Mapping and cartography score	78	69
Recognising infeasible requests score	47	40
Cost per task	$0.354	$0.597
Time per task	8.8 min	10.7 min
Tokens per attempt	289k	627k
Turns per attempt	14.7	16.2
Attempts decided	559	110
Input price per 1M tokens	$1.32	$1.40
Output price per 1M tokens	$3.96	$4.40
Context window	1,048,576	1,048,576
Image input	No	No
Developer	DeepSeek	Z.ai

Charts

Partial coverage. Terminus-2 – DeepSeek V4 Pro (high): 239 of 245 GS-QA2 tasks; Claude Code – GLM-5.3 (max, default): 21 of 147 Earth-Bench V2 tasks; 96 of 173 GeoBenchX tasks; GS-QA2 incomplete. 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. Terminus-2 – DeepSeek V4 Pro (high), Claude Code – GLM-5.3 (max, default).

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. Terminus-2 – DeepSeek V4 Pro (high), Claude Code – GLM-5.3 (max, default).

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 – DeepSeek V4 Pro (high) or Claude Code – GLM-5.3 (max, default)?

Not comparable yet: Terminus-2 – DeepSeek V4 Pro (high) has partial coverage (239 of 245 GS-QA2 tasks) and Claude Code – GLM-5.3 (max, default) has partial coverage (21 of 147 Earth-Bench V2 tasks; 96 of 173 GeoBenchX tasks; GS-QA2 incomplete), so it is not ranked. Terminus-2 – DeepSeek V4 Pro (high): 559 of 859 planned attempts. Claude Code – GLM-5.3 (max, default): 110 of 614 planned attempts. Coverage is partial; scores can change as more attempts are decided.

Which is cheaper per task, Terminus-2 – DeepSeek V4 Pro (high) or Claude Code – GLM-5.3 (max, default)?

Terminus-2 – DeepSeek V4 Pro (high), at $0.354 against $0.597 for Claude Code – GLM-5.3 (max, default).

Which writes fewer output tokens per task, Terminus-2 – DeepSeek V4 Pro (high) or Claude Code – GLM-5.3 (max, default)?

Claude Code – GLM-5.3 (max, default), at 19k against 23k for Terminus-2 – DeepSeek V4 Pro (high).

Which has the larger context window, Terminus-2 – DeepSeek V4 Pro (high) or Claude Code – GLM-5.3 (max, default)?

They are level, at 1,048,576 tokens.

More: Terminus-2 – DeepSeek V4 Pro (high) · Claude Code – GLM-5.3 (max, default) · all comparisons