Laguna XS 2.1 (Reasoning) vs gpt-oss-120b (high)
Results updated 8 November 2026 · In Terminus-2 and Codex, so part of any gap can come from the harness · How we benchmark
Verdict
gpt-oss-120b (high) beats Laguna XS 2.1 (Reasoning) by 6 points, and is faster; Laguna XS 2.1 (Reasoning) is cheaper.
- Cost per task
- $0.054 vs $0.26
- Time per task
- 17.3 min vs 6.0 min
- Tokens per attempt
- 937k vs 700k
- Image input
- No vs No
Laguna XS 2.1 (Reasoning), Terminus-2 scores higher on GS-QA2; gpt-oss-120b (high), Codex scores higher on Earth-Bench V2 and GeoBenchX. Laguna XS 2.1 (Reasoning), Terminus-2 is cheaper ($0.054 against $0.26 per task), but gpt-oss-120b (high), Codex is faster (6.0 min against 17.3 min).
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Laguna XS 2.1 (Reasoning) | gpt-oss-120b (high) |
|---|---|---|
| GIS analysis | 35 | 23 |
| Change detection | 48 | 60 |
| Making maps | 53 | 61 |
| Climate and time series | 36 | 40 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Laguna XS 2.1 (Reasoning), Terminus-2 · gpt-oss-120b (high), Codex.
Side by side
| Metric | Laguna XS 2.1 (Reasoning), Terminus-2 | gpt-oss-120b (high), Codex |
|---|---|---|
| Geospatial Agent Index: Text only | 37 | 43 |
| Geospatial Agent Index: Text + vision | No data | No image input |
| Earth-Bench V2 score | 44 | 49 |
| GeoBenchX score | 33 | 55 |
| GS-QA2 score | 35 | 25 |
| Vision agents score | No data | No image input |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 63 | 58 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 63 | 42 |
| OGD4All score (standalone; in no index) | 65 | No data |
| GeoNatureAgent score (standalone; in no index) | 36 | No data |
| MultiGlobeQA score (standalone; in no index) | 59 | No data |
| GeoAgentEval score (standalone; in no index) | 60 | No data |
| ImpactMesh score (standalone; in no index) | No data | No data |
| Remote sensing and raster analysis score | 39 | 38 |
| Change detection score | 48 | 60 |
| GIS analysis score | 35 | 23 |
| Climate and time series score | 36 | 40 |
| Making maps score | 53 | 61 |
| Cost per task | $0.054 | $0.26 |
| Time per task (at 100 output tokens/s) | 17.3 min | 6.0 min |
| Tokens per attempt | 937k | 700k |
| Turns per attempt | 36.7 | 29.8 |
| Input price per 1M tokens | $0.10 | $0.15 |
| Output price per 1M tokens | $0.20 | $0.59 |
| Context window | 262,144 | 128,000 |
| Image input | No | No |
| Developer | Poolside | OpenAI |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Laguna XS 2.1 (Reasoning), Terminus-2 or gpt-oss-120b (high), Codex?
gpt-oss-120b (high), Codex, 43 to 37.
Which is cheaper per task, Laguna XS 2.1 (Reasoning), Terminus-2 or gpt-oss-120b (high), Codex?
Laguna XS 2.1 (Reasoning), Terminus-2, $0.054 to $0.26.
Which writes fewer output tokens per task, Laguna XS 2.1 (Reasoning), Terminus-2 or gpt-oss-120b (high), Codex?
Laguna XS 2.1 (Reasoning), Terminus-2, 24k to 27k.
Which has the larger context window, Laguna XS 2.1 (Reasoning), Terminus-2 or gpt-oss-120b (high), Codex?
Laguna XS 2.1 (Reasoning), Terminus-2, 262,144 tokens to 128,000 tokens.