// head_to_head
SpaceXAI: Grok 4.20 vs SpaceXAI: Grok 4.3
Published LiveBench scores across all seven categories, live list pricing, context windows, and the measured cost of a point of capability — for both models, side by side.
SpaceXAI: Grok 4.20 and SpaceXAI: Grok 4.3 are priced within ~10% of each other.
SpaceXAI: Grok 4.20 does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
x-ai
SpaceXAI: Grok 4.20
- Blended / 1M
- $1.56
- Context
- 2M
- Released
- Mar 31, 2026
- Overall score
- Not evaluated
x-ai
SpaceXAI: Grok 4.3
- Blended / 1M
- $1.56
- Context
- 1M
- Released
- Apr 30, 2026
- Overall score
- 62.2
Specs and pricing
| Metric | SpaceXAI: Grok 4.20 | SpaceXAI: Grok 4.3 |
|---|---|---|
| LiveBench overall Mean of the seven LiveBench category scores, 0–100. Higher is better. | — | 62.2 |
| Cost per point Measured benchmark spend divided by overall score — dollars per point of capability. | — | $0.0325 |
| Blended price / 1M 3:1 input:output mix, the usual shape of production traffic. | $1.56 | $1.56 |
| Input price / 1M | $1.25 | $1.25 |
| Output price / 1M | $2.50 | $2.50 |
| Cached input / 1M Price of an input token served from the prompt cache, where the provider publishes one. | $0.200 | $0.200 |
| Context window | 2Mwin | 1M |
| Max output tokens | 1.8Mwin | 900K |
Benchmarks by category
An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — SpaceXAI: Grok 4.20 on top, SpaceXAI: Grok 4.3 below, both out of 100.
What each one costs to run
Per-token prices are hard to feel. These are monthly list costs for both models across five workload shapes, using each provider's published cached-input rate where there is one.
| Workload | SpaceXAI: Grok 4.20 | SpaceXAI: Grok 4.3 |
|---|---|---|
| Support chatbot 1.2K in / 400 out × 200K requests | $424.40/mo | $424.40/mo |
| RAG assistant 8K in / 600 out × 100K requests | $730.00/mo | $730.00/mo |
| Coding agent 40K in / 4K out × 20K requests | $612.00/mo | $612.00/mo |
| Document extraction 20K in / 1.5K out × 50K requests | $1,385/mo | $1,385/mo |
| Bulk classification 500 in / 20 out × 5M requests | $2,850/mo | $2,850/mo |
Which should you pick?
You need to fit large documents in one call
SpaceXAI: Grok 4.20
Wider context window — 2M against 1M.
SpaceXAI: Grok 4.20 vs SpaceXAI: Grok 4.3 FAQ
Which is better, SpaceXAI: Grok 4.20 or SpaceXAI: Grok 4.3?
SpaceXAI: Grok 4.20 and SpaceXAI: Grok 4.3 are priced within ~10% of each other. SpaceXAI: Grok 4.20 does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
Is SpaceXAI: Grok 4.20 cheaper than SpaceXAI: Grok 4.3?
They cost about the same. Both land near $1.56 per million tokens on a 3:1 input:output blend, so price is unlikely to be the deciding factor between them.
Does SpaceXAI: Grok 4.20 or SpaceXAI: Grok 4.3 have a bigger context window?
SpaceXAI: Grok 4.20 has the larger context window: 2M for SpaceXAI: Grok 4.20 against 1M for SpaceXAI: Grok 4.3. Note that a window you can fill is not a window you should fill — retrieval quality usually degrades well before the limit, and you pay for every token you put in it.
Do SpaceXAI: Grok 4.20 and SpaceXAI: Grok 4.3 support prompt caching?
Both publish a cached-input rate: $0.200 per million for SpaceXAI: Grok 4.20 and $0.200 for SpaceXAI: Grok 4.3, against full input rates of $1.25 and $1.25. On a workload with a long stable prefix — a system prompt, a tool schema, a retrieved corpus — that changes the economics more than the headline price does.
Related comparisons
How these numbers are produced
- Price — provider list price from OpenRouter, refreshed every 15 minutes. “Blended” is a 3:1 input:output mix.
- Scores — LiveBench release 2026-06-25, using their own category map. Each model shows its strongest published run. A blank means “not evaluated”, never “bad”.
- Cost per point — the measured dollars LiveBench spent on the run, divided by the score it earned.
- “Win” — awarded only past a threshold: one full point on a benchmark score, 10% on a price, 25% on a context window. Anything tighter reports as a tie, because effort settings alone move a LiveBench score by more than that.
Published benchmarks rank models on someone else's tasks. Before committing, see LLM & agent evaluation for building an eval on your own.