Grok 4.20 vs Grok 4.3
Grok 4.20 (xAI) versus Grok 4.3 (xAI): API pricing, context window, capabilities and uptime, from the same data as each model page, at the WandAPI default channel price. Input costs 1.25 Credits / M for Grok 4.20 and 1.25 Credits / M for Grok 4.3. Context windows are 1M and 1M.
Specs side by side
| Item |  Grok 4.20grok-4.20-reasoning | Grok 4.3grok-4.3 |
|---|---|---|
| Input (prompt) · Credits / M | 1.25≈ $0.19 | 1.25≈ $0.19 |
| Output (completion) · Credits / M | 2.5≈ $0.38 | 2.5≈ $0.38 |
| Cache read · Credits / M | 0.2 | 0.2 |
| Context window | 1M | 1M |
| Max output | — | 128K |
| 7-day uptime | — | — |
| Latest probe latency | — | — |
| Measured throughput | 96.2 t/s | — |
| Protocols | OpenAI Responses | OpenAI Responses |
| Sample monthly cost (10M in + 2M out) | 17.5≈ $2.66 | 17.5≈ $2.66 |
Prices are the WandAPI default channel price in Credits (1 Credit = ¥1); green marks the better side. Uptime and latency come from WandAPI probes; a dash means the model is not probed yet.
Capabilities
FAQ
Which has the larger context window, Grok 4.20 or Grok 4.3?
Grok 4.20 offers 1M tokens, Grok 4.3 offers 1M tokens.
Can I switch between Grok 4.20 and Grok 4.3 at any time?
Yes. WandAPI calls every model with the same API key; switching is just changing the model field in the request, and billing is split per model in one console.
How do the capabilities of Grok 4.20 and Grok 4.3 differ?
Their capability flags are identical. Every other capability is either supported by both or by neither; see the capability table above.
More comparisons
One API key for both Grok 4.20 and Grok 4.3
Free credits on sign-up, pay as you go, no monthly fee.