Grok 4.6 vs Grok 4.5
Grok 4.5 is the lower-priced of the two on Router One at base-tier rates (Grok 4.6: request input ≤ 200,000 tokens) — about 27% less on a 1M-input + 1M-output mix. Both carry a 500K context window. Both answer on /v1/chat/completions and /v1/responses (Codex CLI).
Grok 4.6 and Grok 4.5 compared on current per-token rates, context window, and capabilities — both callable through one OpenAI-compatible endpoint with per-request cost traces.
Grok 4.6 vs Grok 4.5: rates, context window, and capabilities
| Spec | Grok 4.6 | Grok 4.5 |
|---|---|---|
| Input / 1M tokens | $1.10 | $0.80 |
| Output / 1M tokens | $3.30 | $2.40 |
| Cached input / 1M tokens | $0.275 | $0.20 |
| Context window | 500K | 500K |
| Capabilities | Chat, Streaming, Tool calling, Vision | Chat, Streaming, Tool calling, Vision |
| Subscription plans | Mid-tier models — Pro, Max, Ultra | Mid-tier models — Pro, Max, Ultra |
| Detail page | Grok 4.6 | Grok 4.5 |
Grok 4.6: request input ≤ 200,000 tokens: Input $1.10 / 1M tokens, Output $3.30 / 1M tokens, Cache write $1.10 / 1M tokens, Cached input $0.275 / 1M tokens; request input > 200,000 tokens: Input $2.20 / 1M tokens, Output $6.60 / 1M tokens, Cache write $2.20 / 1M tokens, Cached input $0.55 / 1M tokens. The total input tokens in each request select the tier; its rates apply to the whole request, not only tokens above the threshold.
What does 1M tokens cost on Grok 4.6 vs Grok 4.5?
Base-tier rates: For a workload of 1M input plus 1M output tokens at current rates: Grok 4.6 comes to $4.40, Grok 4.5 comes to $3.20 — Grok 4.5 is about 27% cheaper on this mix. Real workloads skew heavily toward input tokens, so weigh the input rate by your own ratio; the cached-input row above is the posted catalog rate for that model.
Switch between Grok 4.6 and Grok 4.5 without changing code
Both models are behind the same OpenAI-compatible endpoint, so an A/B test is a one-string change — same key, same code, and every request traced with tokens, cost, and latency in the dashboard:
curl https://api.router.one/v1/chat/completions \
-H "Authorization: Bearer sk-your-router-one-key" \
-H "Content-Type: application/json" \
-d '{"model": "grok-4.6", "messages": [{"role": "user", "content": "Hello"}]}'
# Same request, other model — change one string:
# "model": "grok-4.5"FAQ
Is Grok 4.6 cheaper than Grok 4.5?
Compared at base-tier rates (Grok 4.6: request input ≤ 200,000 tokens). Input: Grok 4.6 $1.10 vs Grok 4.5 $0.80 / 1M tokens; Grok 4.5 has the lower rate. Output: Grok 4.6 $3.30 vs Grok 4.5 $2.40 / 1M tokens; Grok 4.5 has the lower rate. Total cost depends on the workload's input, output and cache usage. Grok 4.6: request input ≤ 200,000 tokens: Input $1.10 / 1M tokens, Output $3.30 / 1M tokens, Cache write $1.10 / 1M tokens, Cached input $0.275 / 1M tokens; request input > 200,000 tokens: Input $2.20 / 1M tokens, Output $6.60 / 1M tokens, Cache write $2.20 / 1M tokens, Cached input $0.55 / 1M tokens. The total input tokens in each request select the tier; its rates apply to the whole request, not only tokens above the threshold. Rates change; the /models page is the live source of truth.
Is Grok 4.6 or Grok 4.5 included in a Router One subscription?
Grok 4.6 and Grok 4.5 both count against the Mid-tier models allowance on Pro, Max and Ultra: Pro 1,200, Max 4,500 and Ultra 15,000 requests per 30-day cycle, one pool shared by every model in that tier. Grok 4.6 requests above its long-context threshold count as more than one quota request. Requests beyond a plan's allowance bill the wallet per token.
Can I switch between Grok 4.6 and Grok 4.5 without changing code?
Yes. Both are served through the same OpenAI-compatible endpoint, so switching is changing the model string in the request — the key, base URL, and request shape stay identical.
Where do these numbers come from?
Specs and prices on this page render from the live Router One catalog — the same data as the /models page — and refresh with it. Pricing methodology is documented on /pricing-methodology. Rates on this page were read from the live catalog on 2026-09-30 (UTC) and refresh within the hour.