DeepSeek V4.1 Flash (max) vs Claude Fable 5.1 (high)
Results updated 3 November 2026 · In Codex and Claude Code, so part of any gap can come from the harness · How we benchmark
Verdict
Claude Fable 5.1 (high) edges DeepSeek V4.1 Flash (max) by 2 points, and is faster; DeepSeek V4.1 Flash (max) is cheaper.
- Cost per task
- $0.067 vs $0.44
- Time per task
- 8.6 min vs 1.1 min
- Tokens per attempt
- 1.17M vs 132k
- Image input
- Yes vs No data
DeepSeek V4.1 Flash (max), Codex scores higher on Earth-Bench V2; Claude Fable 5.1 (high), Claude Code scores higher on GeoBenchX, GS-QA2 and Vision agents. DeepSeek V4.1 Flash (max), Codex is cheaper ($0.067 against $0.44 per task), but Claude Fable 5.1 (high), Claude Code is faster (1.1 min against 8.6 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) | Claude Fable 5.1 (high) |
|---|---|---|
| Change detection | 82 | 64 |
| Remote sensing and raster analysis | 85 | 87 |
| Making maps | 81 | 79 |
| GIS analysis | 90 | 88 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: DeepSeek V4.1 Flash (max), Codex · Claude Fable 5.1 (high), Claude Code.
Side by side
| Metric | DeepSeek V4.1 Flash (max), Codex | Claude Fable 5.1 (high), Claude Code |
|---|---|---|
| Geospatial Agent Index: Text only | 78 | 80 |
| Geospatial Agent Index: Text + vision | 78 | 82 |
| Earth-Bench V2 score | 78 | 72 |
| GeoBenchX score | 66 | 80 |
| GS-QA2 score | 89 | 90 |
| Vision agents score | 80 | 88 |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 96 | 100 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 88 | 95 |
| OGD4All score (standalone; in no index) | No data | 97 |
| GeoNatureAgent score (standalone; in no index) | No data | 100 |
| MultiGlobeQA score (standalone; in no index) | No data | 79 |
| GeoAgentEval score (standalone; in no index) | No data | 100 |
| ImpactMesh score (standalone; in no index) | No data | No data |
| Remote sensing and raster analysis score | 85 | 87 |
| Change detection score | 82 | 64 |
| GIS analysis score | 90 | 88 |
| Climate and time series score | 75 | 76 |
| Making maps score | 81 | 79 |
| Cost per task | $0.067 | $0.44 |
| Time per task (at 100 output tokens/s) | 8.6 min | 1.1 min |
| Tokens per attempt | 1.17M | 132k |
| Turns per attempt | 23.4 | 5.4 |
| Input price per 1M tokens | $0.30 | $10.00 |
| Output price per 1M tokens | $1.20 | $50.00 |
| Context window | 1,048,576 | 200,000 |
| Developer | DeepSeek | Anthropic |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, DeepSeek V4.1 Flash (max), Codex or Claude Fable 5.1 (high), Claude Code?
Claude Fable 5.1 (high), Claude Code, 80 to 78.
Which is cheaper per task, DeepSeek V4.1 Flash (max), Codex or Claude Fable 5.1 (high), Claude Code?
DeepSeek V4.1 Flash (max), Codex, $0.067 to $0.44.
Which writes fewer output tokens per task, DeepSeek V4.1 Flash (max), Codex or Claude Fable 5.1 (high), Claude Code?
Claude Fable 5.1 (high), Claude Code, 3k to 38k.
Which has the larger context window, DeepSeek V4.1 Flash (max), Codex or Claude Fable 5.1 (high), Claude Code?
DeepSeek V4.1 Flash (max), Codex, 1,048,576 tokens to 200,000 tokens.