Claude Fable 5.1 (medium) vs GPT-6.1 Sol (medium)
Results updated 3 November 2026 · In Claude Code and Codex, so part of any gap can come from the harness · How we benchmark
Verdict
GPT-6.1 Sol (medium) edges Claude Fable 5.1 (medium) by 1 point, and is cheaper and faster.
- Cost per task
- $0.33 vs $0.046
- Time per task
- 0.8 min vs 0.4 min
- Tokens per attempt
- 107k vs 89k
- Image input
- No data vs Yes
Claude Fable 5.1 (medium), Claude Code scores higher on Earth-Bench V2 and GeoBenchX; GPT-6.1 Sol (medium), Codex scores higher on GS-QA2. GPT-6.1 Sol (medium), Codex is cheaper and faster: $0.046 against $0.33 per task, and 0.4 min against 0.8 min.
Where they differ most
By kind of work, share of tasks passed, 0 to 100 · Higher is better
| Kind of work | Claude Fable 5.1 (medium) | GPT-6.1 Sol (medium) |
|---|---|---|
| GIS analysis | 86 | 94 |
| Change detection | 72 | 68 |
| Making maps | 84 | 81 |
| Remote sensing and raster analysis | 87 | 85 |
Where the tasks are set: each kind of work links to its tasks and scoresCharts for each: Claude Fable 5.1 (medium), Claude Code · GPT-6.1 Sol (medium), Codex.
Side by side
| Metric | Claude Fable 5.1 (medium), Claude Code | GPT-6.1 Sol (medium), Codex |
|---|---|---|
| Geospatial Agent Index: Text only | 81 | 82 |
| Geospatial Agent Index: Text + vision | 81 | No data |
| Earth-Bench V2 score | 73 | 72 |
| GeoBenchX score | 81 | 79 |
| GS-QA2 score | 89 | 93 |
| Vision agents score | 81 | No data |
| GeoAgentBench score (Guided workflows; not in the Agent Index) | 100 | 100 |
| GeoAnalystBench score (Guided workflows; not in the Agent Index) | 100 | 100 |
| OGD4All score (standalone; in no index) | 97 | No data |
| GeoNatureAgent score (standalone; in no index) | 100 | No data |
| MultiGlobeQA score (standalone; in no index) | 78 | No data |
| GeoAgentEval score (standalone; in no index) | 96 | No data |
| ImpactMesh score (standalone; in no index) | No data | No data |
| Remote sensing and raster analysis score | 87 | 85 |
| Change detection score | 72 | 68 |
| GIS analysis score | 86 | 94 |
| Climate and time series score | 74 | 76 |
| Making maps score | 84 | 81 |
| Cost per task | $0.33 | $0.046 |
| Time per task (at 100 output tokens/s) | 0.8 min | 0.4 min |
| Tokens per attempt | 107k | 89k |
| Turns per attempt | 4.9 | 5.8 |
| Input price per 1M tokens | $10.00 | $2.00 |
| Output price per 1M tokens | $50.00 | $10.00 |
| Context window | 200,000 | 1,050,000 |
| Developer | Anthropic | OpenAI |
Frequently asked questions
Which scores higher on the Geospatial Agent Index, Claude Fable 5.1 (medium), Claude Code or GPT-6.1 Sol (medium), Codex?
GPT-6.1 Sol (medium), Codex, 82 to 81.
Which is cheaper per task, Claude Fable 5.1 (medium), Claude Code or GPT-6.1 Sol (medium), Codex?
GPT-6.1 Sol (medium), Codex, $0.046 to $0.33.
Which writes fewer output tokens per task, Claude Fable 5.1 (medium), Claude Code or GPT-6.1 Sol (medium), Codex?
GPT-6.1 Sol (medium), Codex, 1k to 2k.
Which has the larger context window, Claude Fable 5.1 (medium), Claude Code or GPT-6.1 Sol (medium), Codex?
GPT-6.1 Sol (medium), Codex, 1,050,000 tokens to 200,000 tokens.