Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) vs. Claude Code – GLM-4.7-Flash (thinking on): geospatial agent comparison
Comparison between Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) and Claude Code – GLM-4.7-Flash (thinking on) on the Geospatial Agent Index, each benchmark, cost, time and token use, from the same tasks, agent and limits. Methodology.
Side by side
| Metric | Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) | Claude Code – GLM-4.7-Flash (thinking on) |
|---|---|---|
| Geospatial Agent Index | 47* (partial: attempt 1 on 209 of 565 tasks) | 27* (partial: attempt 1 on 52 of 565 tasks) |
| Earth-Bench V2 score | 44 | 20 |
| GeoBenchX score | 49 | 33 |
| GS-QA2 score | No data | 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 | 46 | 15 |
| Vector and overlay score | 33 | No data |
| Networks and routing score | No data | No data |
| Climate and time series score | 67 | 0 |
| Spatial statistics and interpolation score | 43 | 20 |
| Mapping and cartography score | 73 | 64 |
| Recognising infeasible requests score | 30 | 26 |
| Cost per task | $0.291 | $0.089 |
| Time per task | 3.9 min | 13.6 min |
| Tokens per attempt | 80k | 1.37M |
| Turns per attempt | 7.0 | 35.5 |
| Attempts decided | 209 | 49 |
| Input price per 1M tokens | $3.00 | $0.06 |
| Output price per 1M tokens | $15.00 | $0.40 |
| Context window | 1,000,000 | 131,072 |
| Image input | Yes | No |
| Developer | Anthropic | Z.ai |
Charts
Partial coverage. Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high): 36 of 147 Earth-Bench V2 tasks; GS-QA2 incomplete; Claude Code – GLM-4.7-Flash (thinking on): 10 of 147 Earth-Bench V2 tasks; 42 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. Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high), Claude Code – GLM-4.7-Flash (thinking on).
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. Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high), Claude Code – GLM-4.7-Flash (thinking on).
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, Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) or Claude Code – GLM-4.7-Flash (thinking on)?
Which is cheaper per task, Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) or Claude Code – GLM-4.7-Flash (thinking on)?
Which writes fewer output tokens per task, Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) or Claude Code – GLM-4.7-Flash (thinking on)?
Which has the larger context window, Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) or Claude Code – GLM-4.7-Flash (thinking on)?
More: Terminus-2 – Claude Sonnet 4.6 (adaptive, effort high) · Claude Code – GLM-4.7-Flash (thinking on) · all comparisons