DeepSeek V4 Flash (max) vs Qwen 3.8 27B (Reasoning)
Results updated 14 November 2026 · In Terminus-2 and Qwen Code, so part of any gap can come from the harness · How we benchmark
Verdict
DeepSeek V4 Flash (max) and Qwen 3.8 27B (Reasoning) are level on the Geospatial Agent Index, at 69.
- Cost per task
- $0.062 vs $0.22
- Time per task
- 5.7 min vs 4.7 min (estimated)
- Tokens per attempt
- 303k vs 840k
- Image input
- No vs Yes
DeepSeek V4 Flash (max), Terminus-2 scores higher on GS-QA2; Qwen 3.8 27B (Reasoning), Qwen Code scores higher on Earth-Bench V2 and GeoBenchX. DeepSeek V4 Flash (max), Terminus-2 is cheaper ($0.062 against $0.22 per task), but Qwen 3.8 27B (Reasoning), Qwen Code is faster (4.7 min against 5.7 min).
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | DeepSeek V4 Flash (max) | Qwen 3.8 27B (Reasoning) |
|---|---|---|
| Making maps | 79 | 65 |
| Climate and time series | 66 | 62 |
| Remote sensing and raster analysis | 79 | 77 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: DeepSeek V4 Flash (max), Terminus-2 · Qwen 3.8 27B (Reasoning), Qwen Code.
Side by side
| Metric | DeepSeek V4 Flash (max), Terminus-2 | Qwen 3.8 27B (Reasoning), Qwen Code |
|---|---|---|
| Geospatial Agent Index: Text only | 69 | 69 |
| Geospatial Agent Index: Text + vision | No image input | No data |
| Earth-Bench V2 score | 67 | 69 |
| GeoBenchX score | 62 | 63 |
| GS-QA2 score | 80 | 75 |
| Vision agents score | No image input | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 96 | 94 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 95 | 79 |
| OGD4All score (standalone; in no index) | No data | 97 |
| GeoNatureAgent score (standalone; in no index) | No data | 100 |
| MultiGlobeQA score (standalone; in no index) | No data | 74 |
| 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 | 79 | 77 |
| Change detection score | 72 | 72 |
| GIS analysis score | 75 | 75 |
| Climate and time series score | 66 | 62 |
| Making maps score | 79 | 65 |
| Cost per task | $0.062 | $0.22 |
| Time per task (at 100 output tokens/s) | 5.7 min | 4.7 min (estimated) |
| Tokens per attempt | 303k | 840k |
| Turns per attempt | 14.4 | 17.6 |
| Input price per 1M tokens | $0.15 | $0.45 |
| Output price per 1M tokens | $0.60 | $3.00 |
| Context window | 1,048,576 | 262,144 |
| Image input | No | Yes |
| Developer | DeepSeek | Alibaba (Qwen) |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, DeepSeek V4 Flash (max), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
DeepSeek V4 Flash (max), Terminus-2, 69 to 69.
Which is cheaper per task, DeepSeek V4 Flash (max), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
DeepSeek V4 Flash (max), Terminus-2, $0.062 to $0.22.
Which writes fewer output tokens per task, DeepSeek V4 Flash (max), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
Qwen 3.8 27B (Reasoning), Qwen Code, 28k to 29k.
Which has the larger context window, DeepSeek V4 Flash (max), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
DeepSeek V4 Flash (max), Terminus-2, 1,048,576 tokens to 262,144 tokens.