| GPT-5.6 Sol | Claude Fable 5.2 | |
|---|---|---|
| CTO Index | 67.4 (#19) | 95.5 (#1) |
| Provider | OpenAI | Anthropic |
| Weights | Closed | Closed |
| Price $/M in / out | $4 / $20 | $8 / $40 |
| Context window | Not listed | 1M |
| License | Proprietary | Proprietary |
| Terminal-Bench 4.1 | 38.5% | 65.8% |
| WebDev Arena | 1615 | 1824 |
| ARC-AGI-2 | 92.5% | 93.8% |
| Text Arena | 1485 | 1503 |
| ARC-AGI-3 | 7.8% | 81.2% |
Snapshot 22 November 2026 · Scores copied from each benchmark's own leaderboard · How the CTO Index works
Frequently asked questions
Is GPT-5.6 Sol better than Claude Fable 5.2?
Claude Fable 5.2 scores higher on the CTO Index (95.5 vs 67.4). Across the 5 benchmarks both have results, GPT-5.6 Sol wins 0 and Claude Fable 5.2 wins 5 (Terminal-Bench, WebDev, ARC-AGI-2, Text Arena, ARC-AGI-3). GPT-5.6 Sol costs 50% less per blended token.
Which is better for coding agents, GPT-5.6 Sol or Claude Fable 5.2?
On Terminal-Bench 4.1, Claude Fable 5.2 scores 65.8% against GPT-5.6 Sol's 38.5%. Differences under about three points are within the error bars.
How do GPT-5.6 Sol and Claude Fable 5.2 compare on price?
GPT-5.6 Sol lists at $4 / $20 and Claude Fable 5.2 at $8 / $40 per million input / output tokens.