GLM-5.2 (Non-reasoning) vs Qwen 3.8 27B (Reasoning)
Results updated 24 October 2026 · In Terminus-2 and Qwen Code, so part of any gap can come from the harness · How we benchmark
Verdict
Qwen 3.8 27B (Reasoning) beats GLM-5.2 (Non-reasoning) by 6 points; GLM-5.2 (Non-reasoning) is cheaper and faster.
- Cost per task
- $0.099 vs $0.22
- Time per task
- 1.5 min vs 4.7 min (estimated)
- Tokens per attempt
- 185k vs 840k
- Image input
- No vs Yes
Qwen 3.8 27B (Reasoning), Qwen Code scores higher than GLM-5.2 (Non-reasoning), Terminus-2 on all 3 evaluations scored for both. GLM-5.2 (Non-reasoning), Terminus-2 is cheaper and faster: $0.099 against $0.22 per task, and 1.5 min against 4.7 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | GLM-5.2 (Non-reasoning) | Qwen 3.8 27B (Reasoning) |
|---|---|---|
| Spatial statistics | 31 | 54 |
| Recognising infeasible requests | 45 | 62 |
| Making maps | 79 | 65 |
| GIS analysis | 65 | 75 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: GLM-5.2 (Non-reasoning), Terminus-2 · Qwen 3.8 27B (Reasoning), Qwen Code.
Side by side
| Metric | GLM-5.2 (Non-reasoning), Terminus-2 | Qwen 3.8 27B (Reasoning), Qwen Code |
|---|---|---|
| Geospatial Agent Index: Text only | 63 | 69 |
| Geospatial Agent Index: Text + vision | No image input | No data |
| Earth-Bench V2 score | 58 | 69 |
| GeoBenchX score | 60 | 63 |
| GS-QA2 score | 72 | 75 |
| Vision agents score | No image input | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 92 | 94 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 74 | 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 | 72 | 77 |
| GIS analysis score | 65 | 75 |
| Climate and time series score | 58 | 63 |
| Spatial statistics score | 31 | 54 |
| Making maps score | 79 | 65 |
| Recognising infeasible requests score | 45 | 62 |
| Cost per task | $0.099 | $0.22 |
| Time per task (at 100 output tokens/s) | 1.5 min | 4.7 min (estimated) |
| Tokens per attempt | 185k | 840k |
| Turns per attempt | 12.8 | 17.6 |
| Input price per 1M tokens | $1.40 | $0.45 |
| Output price per 1M tokens | $4.40 | $3.20 |
| Context window | 262,144 | 262,144 |
| Image input | No | Yes |
| Developer | Z.ai | Alibaba (Qwen) |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, GLM-5.2 (Non-reasoning), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
Qwen 3.8 27B (Reasoning), Qwen Code, 69 to 63.
Which is cheaper per task, GLM-5.2 (Non-reasoning), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
GLM-5.2 (Non-reasoning), Terminus-2, $0.099 to $0.22.
Which writes fewer output tokens per task, GLM-5.2 (Non-reasoning), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
GLM-5.2 (Non-reasoning), Terminus-2, 6k to 28k.
Which has the larger context window, GLM-5.2 (Non-reasoning), Terminus-2 or Qwen 3.8 27B (Reasoning), Qwen Code?
They are level, at 262,144 tokens.