Claude Opus 5.5 (medium) vs Claude Haiku 5.5 (max)
Results updated 5 November 2026 · Both in Claude Code · How we benchmark
Verdict
Claude Opus 5.5 (medium) beats Claude Haiku 5.5 (max) by 3 points, and is faster; Claude Haiku 5.5 (max) is cheaper.
- Cost per task
- $0.11 vs $0.088
- Time per task
- 0.5 min vs 8.3 min (estimated)
- Tokens per attempt
- 91k vs 943k
- Image input
- Yes vs Yes
Claude Opus 5.5 (medium), Claude Code scores higher than Claude Haiku 5.5 (max), Claude Code on all 3 evaluations scored for both. Claude Haiku 5.5 (max), Claude Code is cheaper ($0.088 against $0.11 per task), but Claude Opus 5.5 (medium), Claude Code is faster (0.5 min against 8.3 min).
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Claude Opus 5.5 (medium) | Claude Haiku 5.5 (max) |
|---|---|---|
| Change detection | 72 | 56 |
| Remote sensing and raster analysis | 84 | 80 |
| Climate and time series | 74 | 71 |
| Making maps | 86 | 84 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Claude Opus 5.5 (medium), Claude Code · Claude Haiku 5.5 (max), Claude Code.
Side by side
| Metric | Claude Opus 5.5 (medium), Claude Code | Claude Haiku 5.5 (max), Claude Code |
|---|---|---|
| Geospatial Agent Index: Text only | 80 | 77 |
| Geospatial Agent Index: Text + vision | No data | 78 |
| Earth-Bench V2 score | 72 | 67 |
| GeoBenchX score | 79 | 78 |
| GS-QA2 score | 90 | 87 |
| Vision agents score | No data | 81 |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 98 | 98 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 95 | 100 |
| OGD4All score (standalone; in no index) | 97 | 97 |
| GeoNatureAgent score (standalone; in no index) | 100 | 100 |
| MultiGlobeQA score (standalone; in no index) | 78 | 77 |
| GeoAgentEval score (standalone; in no index) | 100 | 96 |
| ImpactMesh score (standalone; in no index) | No data | No data |
| Remote sensing and raster analysis score | 84 | 80 |
| Change detection score | 72 | 56 |
| GIS analysis score | 92 | 90 |
| Climate and time series score | 74 | 71 |
| Making maps score | 86 | 84 |
| Cost per task | $0.11 | $0.088 |
| Time per task (at 100 output tokens/s) | 0.5 min | 8.3 min (estimated) |
| Tokens per attempt | 91k | 943k |
| Turns per attempt | 4.8 | 13.7 |
| Input price per 1M tokens | $4.00 | $0.10 |
| Output price per 1M tokens | $20.00 | $0.50 |
| Context window | 1,000,000 | 1,000,000 |
| Image input | Yes | Yes |
| Developer | Anthropic | Anthropic |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Claude Opus 5.5 (medium), Claude Code or Claude Haiku 5.5 (max), Claude Code?
Claude Opus 5.5 (medium), Claude Code, 80 to 77.
Which is cheaper per task, Claude Opus 5.5 (medium), Claude Code or Claude Haiku 5.5 (max), Claude Code?
Claude Haiku 5.5 (max), Claude Code, $0.088 to $0.11.
Which writes fewer output tokens per task, Claude Opus 5.5 (medium), Claude Code or Claude Haiku 5.5 (max), Claude Code?
Claude Opus 5.5 (medium), Claude Code, 2k to 45k.
Which has the larger context window, Claude Opus 5.5 (medium), Claude Code or Claude Haiku 5.5 (max), Claude Code?
They are level, at 1,000,000 tokens.