Terminus-2 – Gemini 3.1 Flash-Lite (high) vs. Codex – GPT-6 Luna (medium): geospatial agent comparison
Comparison between Terminus-2 – Gemini 3.1 Flash-Lite (high) and Codex – GPT-6 Luna (medium) 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 – Gemini 3.1 Flash-Lite (high) | Codex – GPT-6 Luna (medium) |
|---|---|---|
| Geospatial Agent Index | 56* (partial: attempt 1 on 175 of 488 tasks) | 57* (partial: attempt 1 on 460 of 488 tasks) |
| Earth-Bench V2 score | 52 | 50 |
| GeoBenchX score | 61 | 56 |
| GS-QA2 score | No data | 65 |
| GeoAgentBench score (guided GIS workflow; outside the index) | 69 | 94 |
| GeoAnalystBench score (guided GIS workflow; outside the index) | 58 | 89 |
| Raster and remote sensing score | 52 | 63 |
| Vector and overlay score | 60 | 65 |
| Climate and time series score | 55 | 47 |
| Spatial statistics and interpolation score | 31 | 22 |
| Mapping and cartography score | 69 | 76 |
| Recognising infeasible requests score | 62 | 38 |
| Cost per task | $0.020 | $0.0026 |
| Time per task | 2.5 min | 1.3 min |
| Tokens per attempt | 49k | 86k |
| Turns per attempt | 5.6 | 6.4 |
| Attempts decided | 175 | 460 |
| Input price per 1M tokens | $0.25 | $0.10 |
| Output price per 1M tokens | $1.50 | $0.50 |
| Context window | 1,048,576 | No data |
| Image input | Yes | No data |
| Developer | OpenAI |
Charts
Partial coverage. Terminus-2 – Gemini 3.1 Flash-Lite (high): 27 of 117 Earth-Bench V2 tasks; 148 of 150 GeoBenchX tasks; GS-QA2 incomplete; Codex – GPT-6 Luna (medium): 122 of 150 GeoBenchX tasks. 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. 2 entries without a first attempt on every index task. What is still running is on the models page.
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. 2 entries without a first attempt on every index task. What is still running is on the models page.
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 – Gemini 3.1 Flash-Lite (high) or Codex – GPT-6 Luna (medium)?
Which is cheaper per task, Terminus-2 – Gemini 3.1 Flash-Lite (high) or Codex – GPT-6 Luna (medium)?
Which writes fewer output tokens per task, Terminus-2 – Gemini 3.1 Flash-Lite (high) or Codex – GPT-6 Luna (medium)?
Which has the larger context window, Terminus-2 – Gemini 3.1 Flash-Lite (high) or Codex – GPT-6 Luna (medium)?
More: Terminus-2 – Gemini 3.1 Flash-Lite (high) · Codex – GPT-6 Luna (medium) · all comparisons