Claude Fable 5 vs Kimi K3
Anthropic's Claude Fable 5 and Moonshot AI's Kimi K3 both target production language workloads, but they price and behave differently. Here is the side-by-side.
Pricing checked against provider documentation on . How we verify
The short answer
Kimi K3 is the cheaper option — roughly 3.3x less on a blended workload, and it suits natively multimodal long-horizon agents. Claude Fable 5 justifies its premium when you need long-horizon agentic engineering.
Anthropic • Claude 5
$10.00 / $50.00
/ 1M tokens (input / output)
Anthropic's most capable widely released model, built for agents that run for hours rather than seconds. Adaptive thinking is always on and cannot be disabled, which is part of why it tops the Artificial Analysis Intelligence Index — and why its cost per task runs roughly triple GPT-5.6 Sol's for a one-point intelligence lead.
Moonshot AI • Kimi K
$3.00 / $15.00
/ 1M tokens (input / output)
At 2.8T total parameters with 104B active, Kimi K3 is the largest open-weight model in general circulation and the only one here that ingests video natively. Reasoning is always on with configurable effort. The licence is the catch: it is not a standard open licence, and organisations above $20M in revenue need a separate commercial agreement.
Input price
Kimi K3 is 70% cheaper than Claude Fable 5 on input tokens.
Output price
Kimi K3 is 70% cheaper than Claude Fable 5 on output tokens — usually the side that dominates the bill.
Monthly cost at three workload sizes
Standard (non-batch, non-cached) rates. Reasoning models will exceed these figures because thinking tokens bill as output.
| Workload | Claude Fable 5 | Kimi K3 | Difference |
|---|---|---|---|
| Light — 1M in / 200K out | $20 | $6 | $14 |
| Moderate — 10M in / 2M out | $200 | $60 | $140 |
| Heavy — 100M in / 20M out | $2,000 | $600 | $1,400 |
Specification comparison
| Attribute | Claude Fable 5 | Kimi K3 |
|---|---|---|
| Input (/ 1M tokens) | $10.00 | $3.00 |
| Output (/ 1M tokens) | $50.00 | $15.00 |
| Cached input | $1.00 | $0.30 |
| Context window | 1,000,000 tokens | 1,048,576 tokens |
| Max output | 128,000 tokens | 131,072 tokens |
| Native reasoning | Yes | Yes |
| Knowledge cutoff | 2026-01 | 2026-02 |
| Relative latency | high | high |
| Open weights | No | Kimi K3 License (commercial terms above $20M revenue) |
| API model ID | claude-fable-5 | kimi-k3 |
| Status | stable | stable |
Claude Fable 5: 5-minute cache writes $12.50/MTok, 1-hour $20/MTok. Requires 30-day data retention, so not available under zero-data-retention terms.
Kimi K3: The custom licence requires a separate commercial agreement once the licensee and its affiliates exceed $20M revenue over any consecutive 12 months.
Choose Claude Fable 5 if…
- Multi-hour autonomous coding runs
- Large codebase migrations
- Hard scientific workflows
- High-stakes legal and medical analysis
Choose Kimi K3 if…
- Multimodal agent workflows
- Video and image understanding at length
- Ambitious long-horizon automation
- Research requiring open weights at scale
Frequently asked
Is Claude Fable 5 or Kimi K3 cheaper?
Kimi K3 is cheaper. On a blended 3:1 input-to-output workload it costs about 3.3x less than Claude Fable 5.
Which has the larger context window, Claude Fable 5 or Kimi K3?
Kimi K3 has the larger window at 1.05M tokens versus 1M.
Should I use Claude Fable 5 or Kimi K3?
Pick Claude Fable 5 for long-horizon agentic engineering. Pick Kimi K3 for natively multimodal long-horizon agents. If cost dominates the decision, Kimi K3 wins; if you need the capability ceiling, benchmark both on your own evals before committing.