Axis Spatial

Claude Code – Claude Sonnet 5.5 (high) vs. Claude Code – Claude Haiku 5.5 (xhigh): geospatial agent comparison

Comparison between Claude Code – Claude Sonnet 5.5 (high) and Claude Code – Claude Haiku 5.5 (xhigh) on the Geospatial Agent Index, each benchmark, cost, time and token use, from the same tasks, agent and limits. Methodology.

Side by side

Claude Code – Claude Sonnet 5.5 (high) vs. Claude Code – Claude Haiku 5.5 (xhigh)
MetricClaude Code – Claude Sonnet 5.5 (high)Claude Code – Claude Haiku 5.5 (xhigh)
Geospatial Agent Index81* (partial: 1.9 of 3 attempts)78 (partial: 1 of 3 attempts)
GeoAgentBench score100100
GeoBenchX score6765
Earth-Bench score5852
GeoAnalystBench score9795
Raster and remote sensing score7973
Vector and overlay score8178
Networks and routing score100100
Climate and time series score7971
Spatial statistics and interpolation score7362
Mapping and cartography score7674
Recognising infeasible requests score6165
Cost per task$0.062$0.023
Time per task0.7 min1.9 min
Tokens per attempt95k635k
Turns per attempt4.612.0
Attempts decided479290
Input price per 1M tokens$2.00$0.10
Output price per 1M tokens$10.00$0.50
Context windowNo dataNo data
Image inputNo dataNo data
DeveloperAnthropicAnthropic

Charts

Geospatial Agent Index

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

  • Partial coverage
Geospatial Agent IndexClaude Code – Claude Sonnet 5.5 (high): 81* (partial coverage); Claude Code – Claude Haiku 5.5 (xhigh): 78025507510081*Claude Sonnet 5.5 (high)78Claude Haiku 5.5 (xhigh)
Data table: Geospatial Agent Index
Geospatial Agent Index
ModelCreatorGeospatial Agent IndexRange while attempts are pendingBenchmarks coveredCoverage (attempts)
Claude Code – Claude Sonnet 5.5 (high)Anthropic81* (partial: 1.9 of 3 attempts)71 to 824 of 4479 of 870 planned attempts (236 excluded)
Claude Code – Claude Haiku 5.5 (xhigh)Anthropic78 (partial: 1 of 3 attempts)No pending judgements4 of 4290 of 290 planned attempts

Cost per task

Average cost per task (USD) · Lower is better

  • Partial coverage
Cost per taskClaude Code – Claude Haiku 5.5 (xhigh): $0.023; Claude Code – Claude Sonnet 5.5 (high): $0.062 (partial coverage)$0$0.020$0.040$0.060$0.080$0.023Claude Haiku 5.5 (xhigh)$0.062Claude Sonnet 5.5 (high)
Data table: Cost per task
Cost per task
ModelCreatorCost per taskCoverage (attempts)
Claude Code – Claude Sonnet 5.5 (high)Anthropic$0.062543 of 870 planned attempts (236 excluded)
Claude Code – Claude Haiku 5.5 (xhigh)Anthropic$0.023290 of 290 planned attempts

Time per task

Average agent wall time per task · Lower is better

  • Partial coverage
Time per taskClaude Code – Claude Sonnet 5.5 (high): 0.7 min (partial coverage); Claude Code – Claude Haiku 5.5 (xhigh): 1.9 min0m0m1m2m2m0.7 minClaude Sonnet 5.5 (high)1.9 minClaude Haiku 5.5 (xhigh)
Data table: Time per task
Time per task
ModelCreatorTime per taskCoverage (attempts)
Claude Code – Claude Sonnet 5.5 (high)Anthropic0.7 min543 of 870 planned attempts (236 excluded)
Claude Code – Claude Haiku 5.5 (xhigh)Anthropic1.9 min290 of 290 planned attempts

Frequently asked questions

Which scores higher on the Geospatial Agent Index, Claude Code – Claude Sonnet 5.5 (high) or Claude Code – Claude Haiku 5.5 (xhigh)?
Claude Code – Claude Sonnet 5.5 (high), at 81 against 78 for Claude Code – Claude Haiku 5.5 (xhigh). Claude Code – Claude Sonnet 5.5 (high): 479 of 870 planned attempts (236 excluded). Claude Code – Claude Haiku 5.5 (xhigh): 290 of 290 planned attempts. Coverage is partial; scores can change as more attempts are decided.
Which is cheaper per task, Claude Code – Claude Sonnet 5.5 (high) or Claude Code – Claude Haiku 5.5 (xhigh)?
Claude Code – Claude Haiku 5.5 (xhigh), at $0.023 against $0.062 for Claude Code – Claude Sonnet 5.5 (high).
Which is faster per task, Claude Code – Claude Sonnet 5.5 (high) or Claude Code – Claude Haiku 5.5 (xhigh)?
Claude Code – Claude Sonnet 5.5 (high), at 0.7 min against 1.9 min for Claude Code – Claude Haiku 5.5 (xhigh).
Which has the larger context window, Claude Code – Claude Sonnet 5.5 (high) or Claude Code – Claude Haiku 5.5 (xhigh)?
Not known yet: Claude Code – Claude Sonnet 5.5 (high) has no context window data and Claude Code – Claude Haiku 5.5 (xhigh) has no context window data.

More: Claude Code – Claude Sonnet 5.5 (high) · Claude Code – Claude Haiku 5.5 (xhigh) · all comparisons