Gemini 3.8 Flash (high) vs Gemma 4 31B (Reasoning)
Results updated 11 November 2026 · Both in Terminus-2 · How we benchmark
Verdict
Not comparable yet: Gemini 3.8 Flash (high) is not yet ranked.
- Cost per task
- $0.31 vs $0.015
- Time per task
- 6.7 min vs 2.1 min
- Tokens per attempt
- 994k vs 93k
- Image input
- Yes vs Yes
Gemini 3.8 Flash (high), Terminus-2 scores higher than Gemma 4 31B (Reasoning), Terminus-2 on all 2 evaluations scored for both. Gemma 4 31B (Reasoning), Terminus-2 is cheaper and faster: $0.015 against $0.31 per task, and 2.1 min against 6.7 min.
Not yet ranked. Gemini 3.8 Flash (high), Terminus-2: no GS-QA2 score yet. Not in the Agent Index, cost and time charts until it has a score on every index evaluation. Results so far are in the side-by-side table.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Gemini 3.8 Flash (high) | Gemma 4 31B (Reasoning) |
|---|---|---|
| GIS analysis | 20 | 56 |
| Change detection | 75 | 52 |
| Making maps | 54 | 70 |
| Remote sensing and raster analysis | 48 | 59 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Gemini 3.8 Flash (high), Terminus-2 · Gemma 4 31B (Reasoning), Terminus-2.
Side by side
| Metric | Gemini 3.8 Flash (high), Terminus-2 | Gemma 4 31B (Reasoning), Terminus-2 |
|---|---|---|
| Geospatial Agent Index: Text only | 58 | 55 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 48 | 44 |
| GeoBenchX score | 68 | 61 |
| GS-QA2 score | No data | 61 |
| Vision agents score | No data | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 88 | 83 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 58 | 63 |
| OGD4All score (standalone; in no index) | No data | 85 |
| GeoNatureAgent score (standalone; in no index) | No data | 100 |
| MultiGlobeQA score (standalone; in no index) | No data | 30 |
| GeoAgentEval score (standalone; in no index) | No data | 88 |
| ImpactMesh score (standalone; in no index) | No data | No data |
| Remote sensing and raster analysis score | 48 | 59 |
| Change detection score | 75 | 52 |
| GIS analysis score | 20 | 56 |
| Climate and time series score | 45 | 47 |
| Making maps score | 54 | 70 |
| Cost per task | $0.31 | $0.015 |
| Time per task (at 100 output tokens/s) | 6.7 min | 2.1 min |
| Tokens per attempt | 994k | 93k |
| Turns per attempt | 31.3 | 9.0 |
| Input price per 1M tokens | $0.75 | $0.17 |
| Output price per 1M tokens | $3.75 | $0.40 |
| Context window | 1,048,576 | 262,144 |
| Image input | Yes | Yes |
| Developer |