Kimi K3 (Reasoning) vs Claude Opus 5.5 (max)
Results updated 3 November 2026 · In Kimi Code and Claude Code, so part of any gap can come from the harness · How we benchmark
Verdict
Not comparable yet: Claude Opus 5.5 (max) is not yet ranked.
- Cost per task
- $0.00 vs $0.83
- Time per task
- 1.5 min (estimated) vs 4.5 min (estimated)
- Tokens per attempt
- 345k vs 530k
- Image input
- Yes vs Yes
Kimi K3 (Reasoning), Kimi Code scores higher on Earth-Bench V2; Claude Opus 5.5 (max), Claude Code scores higher on GeoBenchX and GS-QA2. Kimi K3 (Reasoning), Kimi Code is cheaper and faster: $0.00 against $0.83 per task, and 1.5 min against 4.5 min.
Not yet ranked. Claude Opus 5.5 (max), Claude Code. Not in the Agent Index, cost and time charts until it has a score on every index evaluation. Results so far are in the side-by-side table.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Kimi K3 (Reasoning) | Claude Opus 5.5 (max) |
|---|---|---|
| Recognising infeasible requests | 52 | 79 |
| Climate and time series | 78 | 60 |
| Spatial statistics | 38 | 50 |
| GIS analysis | 84 | 93 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Kimi K3 (Reasoning), Kimi Code · Claude Opus 5.5 (max), Claude Code.
Side by side
| Metric | Kimi K3 (Reasoning), Kimi Code | Claude Opus 5.5 (max), Claude Code |
|---|---|---|
| Geospatial Agent Index: Text only | 75 | 79 |
| Geospatial Agent Index: Text + vision | No data | No data |
| Earth-Bench V2 score | 74 | 71 |
| GeoBenchX score | 65 | 74 |
| GS-QA2 score | 85 | 92 |
| Vision agents score | No data | 98 |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 96 | 100 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 100 | 100 |
| OGD4All score (standalone; in no index) | 97 | 95 |
| GeoNatureAgent score (standalone; in no index) | 100 | 100 |
| MultiGlobeQA score (standalone; in no index) | 77 | 76 |
| GeoAgentEval score (standalone; in no index) | 96 | 100 |
| Remote sensing and raster analysis score | 81 | 88 |
| GIS analysis score | 84 | 93 |
| Climate and time series score | 78 | 60 |
| Spatial statistics score | 38 | 50 |
| Making maps score | 82 | 77 |
| Recognising infeasible requests score | 52 | 79 |
| Cost per task | $0.00 | $0.83 |
| Time per task (at 100 output tokens/s) | 1.5 min (estimated) | 4.5 min (estimated) |
| Tokens per attempt | 345k | 530k |
| Turns per attempt | No data | 12.0 |
| Input price per 1M tokens | $3.00 | $4.00 |
| Output price per 1M tokens | $15.00 | $20.00 |
| Context window | 1,000,000 | 1,000,000 |
| Image input | Yes | Yes |
| Developer | Moonshot AI | Anthropic |