MiniMax M3.1 Flash (max) vs Mistral Large 4 (high)
Results updated 3 November 2026 · Both in Terminus-2 · How we benchmark
Verdict
Mistral Large 4 (high) beats MiniMax M3.1 Flash (max) by 7 points; MiniMax M3.1 Flash (max) is cheaper and faster.
- Cost per task
- $0.018 vs $0.10
- Time per task
- 8.7 min vs 12.0 min
- Tokens per attempt
- 263k vs 226k
- Image input
- Yes vs Yes
Mistral Large 4 (high), Terminus-2 scores higher than MiniMax M3.1 Flash (max), Terminus-2 on all 3 evaluations scored for both. MiniMax M3.1 Flash (max), Terminus-2 is cheaper and faster: $0.018 against $0.10 per task, and 8.7 min against 12.0 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | MiniMax M3.1 Flash (max) | Mistral Large 4 (high) |
|---|---|---|
| Spatial statistics | 23 | 46 |
| Making maps | 60 | 82 |
| Remote sensing and raster analysis | 57 | 67 |
| Recognising infeasible requests | 32 | 23 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: MiniMax M3.1 Flash (max), Terminus-2 · Mistral Large 4 (high), Terminus-2.
Side by side
| Metric | MiniMax M3.1 Flash (max), Terminus-2 | Mistral Large 4 (high), Terminus-2 |
|---|---|---|
| Geospatial Agent Index: Text only | 52 | 59 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 39 | 46 |
| GeoBenchX score | 44 | 55 |
| GS-QA2 score | 71 | 77 |
| Vision agents score | No data | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 88 | 80 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 47 | 94 |
| OGD4All score (standalone; in no index) | 88 | 96 |
| GeoNatureAgent score (standalone; in no index) | 100 | 100 |
| MultiGlobeQA score (standalone; in no index) | 65 | 62 |
| GeoAgentEval score (standalone; in no index) | 67 | 60 |
| Remote sensing and raster analysis score | 57 | 67 |
| GIS analysis score | 77 | 82 |
| Climate and time series score | 34 | 40 |
| Spatial statistics score | 23 | 46 |
| Making maps score | 60 | 82 |
| Recognising infeasible requests score | 32 | 23 |
| Cost per task | $0.018 | $0.10 |
| Time per task (at 100 output tokens/s) | 8.7 min | 12.0 min |
| Tokens per attempt | 263k | 226k |
| Turns per attempt | 13.3 | 13.8 |
| Input price per 1M tokens | $0.15 | $0.70 |
| Output price per 1M tokens | $0.60 | $2.10 |
| Context window | 1,000,000 | 262,144 |
| Image input | Yes | Yes |
| Developer | MiniMax | Mistral AI |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, MiniMax M3.1 Flash (max), Terminus-2 or Mistral Large 4 (high), Terminus-2?
Mistral Large 4 (high), Terminus-2, 59 to 52.
Which is cheaper per task, MiniMax M3.1 Flash (max), Terminus-2 or Mistral Large 4 (high), Terminus-2?
MiniMax M3.1 Flash (max), Terminus-2, $0.018 to $0.10.
Which writes fewer output tokens per task, MiniMax M3.1 Flash (max), Terminus-2 or Mistral Large 4 (high), Terminus-2?
Mistral Large 4 (high), Terminus-2, 14k to 19k.
Which has the larger context window, MiniMax M3.1 Flash (max), Terminus-2 or Mistral Large 4 (high), Terminus-2?
MiniMax M3.1 Flash (max), Terminus-2, 1,000,000 tokens to 262,144 tokens.