Claude Fable 5
Claude Fable 5 is developed by Anthropic. MeetNewAI lists 6 benchmark boards and 4 plans for it.
Plans and real price
Cheapest measured$0.108per million tokens, measuredClaude Max 5x (9/14+) · low confidence · Aug 1, 2026
API list price$10 / $50input / output per million tokens
| Plan | Monthly price | Usable tokens / month | Real price / MTok | Confidence | Data date |
|---|
| Claude Max 5x (9/14+) | $100 | 923.5M | $0.108 | low | Aug 1, 2026 |
| Claude Max 20x (9/14+) | $200 | 1.2B | $0.166 | low | Aug 1, 2026 |
| Droid Max | $200 | 1B | $0.198 | medium | Sep 29, 2026 |
| Claude Fable 5 API | — | — | $1.53 | high | Sep 5, 2026 |
Data: real-api-pricing · Sources and licenses
Benchmarks
| Configuration | Harness | Score | Cost per task |
|---|
| Claude Fable 5 (with fallback) | — | 49.6 | — |
| Configuration | Harness | Score | Cost per task |
|---|
| Claude Code - Fable 5 (max) | Claude Code | 44.5 (40.7–48.4) | — |
| Configuration | Harness | Score | Cost per task |
|---|
| Artificial Analysis - Claude Fable 5 (with fallback) | Artificial Analysis | 42.4 | — |
| Configuration | Harness | Score | Cost per task |
|---|
| claude-fable-5-high | — | 1626.0 (1619.6–1632.4) | — |
| Configuration | Harness | Score | Cost per task |
|---|
| Claude Fable 5 (High) | — | 8.4 (7.1–9.7) | — |
| Configuration | Harness | Score | Cost per task |
|---|
| mini-swe-agent - claude-fable-5 (xhigh) | mini-swe-agent | 69.9 (66.7–73.2) | $13 |
| mini-swe-agent - claude-fable-5 (max) | mini-swe-agent | 69.7 (65.7–73.8) | $22 |
| mini-swe-agent - claude-fable-5 (high) | mini-swe-agent | 68.6 (67.5–69.7) | $9.18 |
| mini-swe-agent - claude-fable-5 (medium) | mini-swe-agent | 65.4 (60.9–69.8) | $6.09 |
| mini-swe-agent - claude-fable-5 (low) | mini-swe-agent | 59.6 (56.8–62.4) | $3.76 |
Data: real-api-pricing · Sources and licenses
