DeepSeek V4.1 Flash (max) vs MiniMax M3 (Reasoning)
Results updated 1 November 2026 · In Claude Code and MiniMax Code, so part of any gap can come from the harness · How we benchmark
Verdict
DeepSeek V4.1 Flash (max) and MiniMax M3 (Reasoning) are level on the Geospatial Agent Index, at 68.
- Cost per task
- $0.20 vs $0.15
- Time per task
- 10.9 min vs 4.9 min (estimated)
- Tokens per attempt
- 1.58M vs 1.68M
- Image input
- Yes vs Yes
DeepSeek V4.1 Flash (max), Claude Code scores higher on GS-QA2; MiniMax M3 (Reasoning), MiniMax Code scores higher on Earth-Bench V2 and GeoBenchX. MiniMax M3 (Reasoning), MiniMax Code is cheaper and faster: $0.15 against $0.20 per task, and 4.9 min against 10.9 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | DeepSeek V4.1 Flash (max) | MiniMax M3 (Reasoning) |
|---|---|---|
| Spatial statistics | 31 | 62 |
| GIS analysis | 92 | 81 |
| Climate and time series | 64 | 74 |
| Recognising infeasible requests | 40 | 33 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: DeepSeek V4.1 Flash (max), Claude Code · MiniMax M3 (Reasoning), MiniMax Code.
Side by side
| Metric | DeepSeek V4.1 Flash (max), Claude Code | MiniMax M3 (Reasoning), MiniMax Code |
|---|---|---|
| Geospatial Agent Index: Text only | 68 | 68 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 58 | 68 |
| GeoBenchX score | 55 | 59 |
| GS-QA2 score | 89 | 78 |
| Vision agents score | No data | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 96 | 94 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 100 | 79 |
| OGD4All score (standalone; in no index) | 79 | 94 |
| GeoNatureAgent score (standalone; in no index) | 100 | 93 |
| MultiGlobeQA score (standalone; in no index) | 64 | 73 |
| GeoAgentEval score (standalone; in no index) | 84 | 92 |
| Remote sensing and raster analysis score | 75 | 73 |
| GIS analysis score | 92 | 81 |
| Climate and time series score | 64 | 74 |
| Spatial statistics score | 31 | 62 |
| Making maps score | 72 | 77 |
| Recognising infeasible requests score | 40 | 33 |
| Cost per task | $0.20 | $0.15 |
| Time per task (at 100 output tokens/s) | 10.9 min | 4.9 min (estimated) |
| Tokens per attempt | 1.58M | 1.68M |
| Turns per attempt | 26.0 | No data |
| Input price per 1M tokens | $0.30 | $0.30 |
| Output price per 1M tokens | $1.20 | $1.20 |
| Context window | 1,048,576 | 1,048,576 |
| Image input | Yes | Yes |
| Developer | DeepSeek | MiniMax |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, DeepSeek V4.1 Flash (max), Claude Code or MiniMax M3 (Reasoning), MiniMax Code?
MiniMax M3 (Reasoning), MiniMax Code, 68 to 68.
Which is cheaper per task, DeepSeek V4.1 Flash (max), Claude Code or MiniMax M3 (Reasoning), MiniMax Code?
MiniMax M3 (Reasoning), MiniMax Code, $0.15 to $0.20.
Which writes fewer output tokens per task, DeepSeek V4.1 Flash (max), Claude Code or MiniMax M3 (Reasoning), MiniMax Code?
MiniMax M3 (Reasoning), MiniMax Code, 29k to 48k.
Which has the larger context window, DeepSeek V4.1 Flash (max), Claude Code or MiniMax M3 (Reasoning), MiniMax Code?
They are level, at 1,048,576 tokens.