Claude Code – Claude Haiku 4.5 (thinking 32k) vs. Claude Code – GLM-5.3 (max, default): geospatial agent comparison
Comparison between Claude Code – Claude Haiku 4.5 (thinking 32k) 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
| Metric | Claude Code – Claude Haiku 4.5 (thinking 32k) | Claude Code – GLM-5.3 (max, default) |
|---|---|---|
| Geospatial Agent Index | 49* (partial: attempt 1 on 453 of 565 tasks) | 64* (partial: attempt 1 on 117 of 565 tasks) |
| Earth-Bench V2 score | 36 | 76 |
| GeoBenchX score | 46 | 52 |
| GS-QA2 score | 65 | No data |
| GeoAgentBench score (guided GIS workflow; outside the index) | 79 | No data |
| GeoAnalystBench score (guided GIS workflow; outside the index) | 80 | No data |
| Raster and remote sensing score | 58 | 68 |
| Vector and overlay score | 64 | 67 |
| Networks and routing score | No data | No data |
| Climate and time series score | 33 | 100 |
| Spatial statistics and interpolation score | 25 | 33 |
| Mapping and cartography score | 56 | 69 |
| Recognising infeasible requests score | 36 | 40 |
| Cost per task | $0.099 | $0.597 |
| Time per task | 1.6 min | 10.7 min |
| Tokens per attempt | 333k | 627k |
| Turns per attempt | 11.0 | 16.2 |
| Attempts decided | 453 | 110 |
| Input price per 1M tokens | $1.00 | $1.40 |
| Output price per 1M tokens | $5.00 | $4.40 |
| Context window | 200,000 | 1,048,576 |
| Image input | Yes | No |
| Developer | Anthropic | Z.ai |
Charts
Partial coverage. Claude Code – Claude Haiku 4.5 (thinking 32k): 102 of 147 Earth-Bench V2 tasks; 106 of 173 GeoBenchX 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
First eight selected entries in metric order. Full results in table view.
Data table: 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
First eight selected entries in metric order. Full results in table view.
How to read this chart
Not shown: partial coverage. Claude Code – Claude Haiku 4.5 (thinking 32k), Claude Code – GLM-5.3 (max, default).
Data table: Cost per task
| Model | Creator | Cost per task | Coverage (attempts) |
|---|
Output tokens per task
Average output tokens per attempt, reasoning included · Lower is better
First eight selected entries in metric order. Full results in table view.
How to read this chart
Not shown: partial coverage. Claude Code – Claude Haiku 4.5 (thinking 32k), Claude Code – GLM-5.3 (max, default).
Data table: Output tokens per task
| Model | Creator | Output tokens per task | Coverage (attempts) |
|---|
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Claude Code – Claude Haiku 4.5 (thinking 32k) or Claude Code – GLM-5.3 (max, default)?
Which is cheaper per task, Claude Code – Claude Haiku 4.5 (thinking 32k) or Claude Code – GLM-5.3 (max, default)?
Which writes fewer output tokens per task, Claude Code – Claude Haiku 4.5 (thinking 32k) or Claude Code – GLM-5.3 (max, default)?
Which has the larger context window, Claude Code – Claude Haiku 4.5 (thinking 32k) or Claude Code – GLM-5.3 (max, default)?
More: Claude Code – Claude Haiku 4.5 (thinking 32k) · Claude Code – GLM-5.3 (max, default) · all comparisons