Terminus-2 – Gemini 3.8 Flash (low) vs. Codex – GPT-6 Luna (xhigh): geospatial agent comparison
Comparison between Terminus-2 – Gemini 3.8 Flash (low) and Codex – GPT-6 Luna (xhigh) 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 3.8 Flash (low) | Codex – GPT-6 Luna (xhigh) |
|---|---|---|
| Geospatial Agent Index | 64* (attempts per task: 2.9) | 69* (attempts per task: 3.3) |
| GeoAgentBench score | 96 | 100 |
| GeoBenchX score | 61 | 56 |
| Earth-Bench score | 37 | 18 |
| GeoAnalystBench score | 63 | 100 |
| Raster and remote sensing score | 62 | 50 |
| Vector and overlay score | 63 | 78 |
| Networks and routing score | 100 | 100 |
| Climate and time series score | 51 | 36 |
| Spatial statistics and interpolation score | 43 | 50 |
| Mapping and cartography score | 53 | 66 |
| Recognising infeasible requests score | 83 | 61 |
| Cost per task | $0.115 | $0.0095 |
| Time per task | 2.7 min | 3.5 min |
| Tokens per attempt | 314k | 238k |
| Turns per attempt | 17.9 | 10.0 |
| Attempts decided | 833 | 444 |
| Input price per 1M tokens | $0.75 | $0.10 |
| Output price per 1M tokens | $3.75 | $0.50 |
| Context window | 1,048,576 | No data |
| Image input | Yes | No data |
| Developer | OpenAI |
Charts
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) |
|---|---|---|---|---|---|
| Codex – GPT-6 Luna (xhigh) | OpenAI | 69* (attempts per task: 3.3) | 34 to 86 | 4 of 4 | 444 of 1,167 planned attempts |
| Terminus-2 – Gemini 3.8 Flash (low) | 64* (attempts per task: 2.9) | No pending judgements; coverage incomplete | 4 of 4 | 833 of 1,167 planned attempts (74 excluded) |
Codex – GPT-6 Luna (xhigh)
Terminus-2 – Gemini 3.8 Flash (low)Cost per task
Average cost per task (USD) · Lower is better
First eight selected entries in metric order. Full results in table view.
Data table: Cost per task
| Model | Creator | Cost per task | Coverage (attempts) |
|---|---|---|---|
| Terminus-2 – Gemini 3.8 Flash (low) | $0.115 | 833 of 1,167 planned attempts (74 excluded) | |
| Codex – GPT-6 Luna (xhigh) | OpenAI | $0.0095 | 945 of 1,167 planned attempts |
Time per task
Average agent wall time per task · Lower is better
First eight selected entries in metric order. Full results in table view.
Data table: Time per task
| Model | Creator | Time per task | Coverage (attempts) |
|---|---|---|---|
| Terminus-2 – Gemini 3.8 Flash (low) | 2.7 min | 833 of 1,167 planned attempts (74 excluded) | |
| Codex – GPT-6 Luna (xhigh) | OpenAI | 3.5 min | 945 of 1,167 planned attempts |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Terminus-2 – Gemini 3.8 Flash (low) or Codex – GPT-6 Luna (xhigh)?
Codex – GPT-6 Luna (xhigh), at 69 against 64 for Terminus-2 – Gemini 3.8 Flash (low). Terminus-2 – Gemini 3.8 Flash (low): 833 of 1,167 planned attempts (74 excluded). Codex – GPT-6 Luna (xhigh): 444 of 1,167 planned attempts. Coverage is partial; scores can change as more attempts are decided.
Which is cheaper per task, Terminus-2 – Gemini 3.8 Flash (low) or Codex – GPT-6 Luna (xhigh)?
Codex – GPT-6 Luna (xhigh), at $0.0095 against $0.115 for Terminus-2 – Gemini 3.8 Flash (low).
Which is faster per task, Terminus-2 – Gemini 3.8 Flash (low) or Codex – GPT-6 Luna (xhigh)?
Terminus-2 – Gemini 3.8 Flash (low), at 2.7 min against 3.5 min for Codex – GPT-6 Luna (xhigh).
Which has the larger context window, Terminus-2 – Gemini 3.8 Flash (low) or Codex – GPT-6 Luna (xhigh)?
Not known yet: Codex – GPT-6 Luna (xhigh) has no context window data.
More: Terminus-2 – Gemini 3.8 Flash (low) · Codex – GPT-6 Luna (xhigh) · all comparisons