Kimi K2.7 Code (Reasoning) vs MiniMax M3 (Reasoning)
Results updated 1 November 2026 · In Terminus-2 and MiniMax Code, so part of any gap can come from the harness · How we benchmark
Verdict
Kimi K2.7 Code (Reasoning) beats MiniMax M3 (Reasoning) by 3 points, and is cheaper and faster.
- Cost per task
- $0.11 vs $0.15
- Time per task
- 4.0 min vs 4.9 min (estimated)
- Tokens per attempt
- 273k vs 1.68M
- Image input
- Yes vs Yes
Kimi K2.7 Code (Reasoning), Terminus-2 scores higher than MiniMax M3 (Reasoning), MiniMax Code on all 3 evaluations scored for both. Kimi K2.7 Code (Reasoning), Terminus-2 is cheaper and faster: $0.11 against $0.15 per task, and 4.0 min against 4.9 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Kimi K2.7 Code (Reasoning) | MiniMax M3 (Reasoning) |
|---|---|---|
| Recognising infeasible requests | 57 | 33 |
| Spatial statistics | 38 | 62 |
| Making maps | 74 | 77 |
| Remote sensing and raster analysis | 76 | 73 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Kimi K2.7 Code (Reasoning), Terminus-2 · MiniMax M3 (Reasoning), MiniMax Code.
Side by side
| Metric | Kimi K2.7 Code (Reasoning), Terminus-2 | MiniMax M3 (Reasoning), MiniMax Code |
|---|---|---|
| Geospatial Agent Index: Text only | 71 | 68 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 69 | 68 |
| GeoBenchX score | 62 | 59 |
| GS-QA2 score | 82 | 78 |
| Vision agents score | No data | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 92 | 94 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 89 | 79 |
| OGD4All score (standalone; in no index) | No data | 94 |
| GeoNatureAgent score (standalone; in no index) | No data | 93 |
| MultiGlobeQA score (standalone; in no index) | No data | 73 |
| GeoAgentEval score (standalone; in no index) | No data | 92 |
| Remote sensing and raster analysis score | 76 | 73 |
| GIS analysis score | 81 | 81 |
| Climate and time series score | 75 | 74 |
| Spatial statistics score | 38 | 62 |
| Making maps score | 74 | 77 |
| Recognising infeasible requests score | 57 | 33 |
| Cost per task | $0.11 | $0.15 |
| Time per task (at 100 output tokens/s) | 4.0 min | 4.9 min (estimated) |
| Tokens per attempt | 273k | 1.68M |
| Turns per attempt | 14.2 | No data |
| Input price per 1M tokens | $0.95 | $0.30 |
| Output price per 1M tokens | $4.00 | $1.20 |
| Context window | 262,144 | 1,048,576 |
| Image input | Yes | Yes |
| Developer | Moonshot AI | MiniMax |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Kimi K2.7 Code (Reasoning), Terminus-2 or MiniMax M3 (Reasoning), MiniMax Code?
Kimi K2.7 Code (Reasoning), Terminus-2, 71 to 68.
Which is cheaper per task, Kimi K2.7 Code (Reasoning), Terminus-2 or MiniMax M3 (Reasoning), MiniMax Code?
Kimi K2.7 Code (Reasoning), Terminus-2, $0.11 to $0.15.
Which writes fewer output tokens per task, Kimi K2.7 Code (Reasoning), Terminus-2 or MiniMax M3 (Reasoning), MiniMax Code?
Kimi K2.7 Code (Reasoning), Terminus-2, 12k to 29k.
Which has the larger context window, Kimi K2.7 Code (Reasoning), Terminus-2 or MiniMax M3 (Reasoning), MiniMax Code?
MiniMax M3 (Reasoning), MiniMax Code, 1,048,576 tokens to 262,144 tokens.