Claude Haiku 4.5 vs Grok 4.6
Claude Haiku 4.5 is the cheaper of the two across every workload shape tested below. Headline rates are $1 in / $5 out for Claude Haiku 4.5 versus $2 in / $6 out for Grok 4.6, per million tokens.
Real monthly cost, three workloads
| Workload | Claude Haiku 4.5 | Grok 4.6 | Cheaper |
|---|---|---|---|
| Support chatbot 1.5k in / 400 out · 100k req/mo · 30% cached | $309.50 | $472.50 | Claude Haiku 4.5 (1.5× cheaper) |
| RAG / doc Q&A 12k in / 700 out · 50k req/mo · 60% cached | $451.00 | $870.00 | Claude Haiku 4.5 (1.9× cheaper) |
| Coding agent 25k in / 2.5k out · 20k req/mo · 80% cached | $390.00 | $700.00 | Claude Haiku 4.5 (1.8× cheaper) |
Costs include prompt caching at the stated hit rate and per-model tokenizer correction. Adjust for your own traffic →
Specifications side by side
| Claude Haiku 4.5 | Grok 4.6 | |
|---|---|---|
| Provider | Anthropic | xAI |
| Input / M tokens | $1 | $2 |
| Output / M tokens | $5 | $6 |
| Cached input | $0.1 | $0.5 |
| Context window | 200K | 500K |
| Batch discount | −50% | — |
| Tokenizer factor | ×1 | ×1 |
| Min prefix to cache | 4096 | — |
| Announced retirement | 2026-10-15+ | — |
Which should you pick?
On cost alone, Claude Haiku 4.5 wins every workload shape above.
Caching behaviour differs. Claude Haiku 4.5 requires a prefix of at least 4,096 tokens before caching activates.
Below those thresholds cache_control is ignored silently, with no error — which can quietly
erase the saving you were counting on.
Lifecycle matters as much as price here. Claude Haiku 4.5 has an announced shutdown date of 2026-10-15 or later. A model that is marginally cheaper but retires within your planning horizon is rarely the better choice.
Full detail: Claude Haiku 4.5 · Grok 4.6
All figures verified 2026-08-18 against official provider documentation. Fields a provider does not publish are shown as “—” rather than estimated.