Qwen 3.8 27B (xhigh) vs MiMo V2.6 Flash (Reasoning)
Results updated 1 November 2026 · In Terminus-2 and MiMo Code, so part of any gap can come from the harness · How we benchmark
Verdict
Qwen 3.8 27B (xhigh) edges MiMo V2.6 Flash (Reasoning) by 2 points; MiMo V2.6 Flash (Reasoning) is cheaper and faster.
- Cost per task
- $0.12 vs $0.047
- Time per task
- 7.4 min vs 5.3 min (estimated)
- Tokens per attempt
- 396k vs 2.16M
- Image input
- Yes vs Yes
Qwen 3.8 27B (xhigh), Terminus-2 scores higher on GeoBenchX; MiMo V2.6 Flash (Reasoning), MiMo Code scores higher on Earth-Bench V2 and GS-QA2. MiMo V2.6 Flash (Reasoning), MiMo Code is cheaper and faster: $0.047 against $0.12 per task, and 5.3 min against 7.4 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Qwen 3.8 27B (xhigh) | MiMo V2.6 Flash (Reasoning) |
|---|---|---|
| Spatial statistics | 69 | 31 |
| Recognising infeasible requests | 57 | 48 |
| Making maps | 70 | 67 |
| Remote sensing and raster analysis | 71 | 73 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Qwen 3.8 27B (xhigh), Terminus-2 · MiMo V2.6 Flash (Reasoning), MiMo Code.
Side by side
| Metric | Qwen 3.8 27B (xhigh), Terminus-2 | MiMo V2.6 Flash (Reasoning), MiMo Code |
|---|---|---|
| Geospatial Agent Index: Text only | 66 | 64 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 51 | 55 |
| GeoBenchX score | 67 | 57 |
| GS-QA2 score | 79 | 82 |
| Vision agents score | No data | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 62 | 90 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 89 | 79 |
| OGD4All score (standalone; in no index) | 92 | 84 |
| GeoNatureAgent score (standalone; in no index) | 100 | 92 |
| MultiGlobeQA score (standalone; in no index) | 71 | 56 |
| GeoAgentEval score (standalone; in no index) | 92 | 76 |
| Remote sensing and raster analysis score | 71 | 73 |
| GIS analysis score | 81 | 83 |
| Climate and time series score | 49 | 51 |
| Spatial statistics score | 69 | 31 |
| Making maps score | 70 | 67 |
| Recognising infeasible requests score | 57 | 48 |
| Cost per task | $0.12 | $0.047 |
| Time per task (at 100 output tokens/s) | 7.4 min | 5.3 min (estimated) |
| Tokens per attempt | 396k | 2.16M |
| Turns per attempt | 15.7 | 23.6 |
| Input price per 1M tokens | $0.45 | $0.14 |
| Output price per 1M tokens | $3.20 | $0.28 |
| Context window | 262,144 | 1,048,576 |
| Image input | Yes | Yes |
| Developer | Alibaba (Qwen) | Xiaomi |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Qwen 3.8 27B (xhigh), Terminus-2 or MiMo V2.6 Flash (Reasoning), MiMo Code?
Qwen 3.8 27B (xhigh), Terminus-2, 66 to 64.
Which is cheaper per task, Qwen 3.8 27B (xhigh), Terminus-2 or MiMo V2.6 Flash (Reasoning), MiMo Code?
MiMo V2.6 Flash (Reasoning), MiMo Code, $0.047 to $0.12.
Which writes fewer output tokens per task, Qwen 3.8 27B (xhigh), Terminus-2 or MiMo V2.6 Flash (Reasoning), MiMo Code?
Qwen 3.8 27B (xhigh), Terminus-2, 22k to 32k.
Which has the larger context window, Qwen 3.8 27B (xhigh), Terminus-2 or MiMo V2.6 Flash (Reasoning), MiMo Code?
MiMo V2.6 Flash (Reasoning), MiMo Code, 1,048,576 tokens to 262,144 tokens.