// head_to_head

DeepSeek V4 Pro 0423 vs SpaceXAI: Grok 4.5

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.

OpenRouter + LiveBenchAll comparisonsFull leaderboard

SpaceXAI: Grok 4.5 scores higher, DeepSeek V4 Pro 0423 costs less — it depends on your workload.

SpaceXAI: Grok 4.5 is ahead by 4.2 points overall, and DeepSeek V4 Pro 0423 lists 4.5× cheaper per blended million tokens. Whether 4.2 points is worth that depends on how much a wrong answer costs you. DeepSeek V4 Pro 0423 also leads on measured cost per point of capability, at $0.0261 per point.

deepseek

DeepSeek V4 Pro 0423

Blended / 1M
$0.665
Context
1.0M
Released
Apr 24, 2026
Overall score
71.6
reasoningtool callingprompt caching

x-ai

SpaceXAI: Grok 4.5

Blended / 1M
$3.00
Context
500K
Released
Jul 8, 2026
Overall score
75.8
reasoningtool callingimage inputfile inputprompt caching

Specs and pricing

MetricDeepSeek V4 Pro 0423SpaceXAI: Grok 4.5
LiveBench overall

Mean of the seven LiveBench category scores, 0–100. Higher is better.

71.675.8win
Cost per point

Measured benchmark spend divided by overall score — dollars per point of capability.

$0.0261win$0.0720
Blended price / 1M

3:1 input:output mix, the usual shape of production traffic.

$0.665win$3.00
Input price / 1M$0.532win$2.00
Output price / 1M$1.06win$6.00
Cached input / 1M

Price of an input token served from the prompt cache, where the provider publishes one.

$0.044win$0.300
Context window1.0Mwin500K
Max output tokens384K

Benchmarks by category

An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — DeepSeek V4 Pro 0423 on top, SpaceXAI: Grok 4.5 below, both out of 100.

Agentic coding
42.6
56.5
Coding
70.0
68.6
Reasoning
82.7
87.2
Mathematicstoo close to call
90.7
90.8
Data analysis
74.5
73.0
Language
78.1
82.8
Instruction following
62.4
71.5

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.

WorkloadDeepSeek V4 Pro 0423SpaceXAI: Grok 4.5
Support chatbot

1.2K in / 400 out × 200K requests

$177.72/mo$837.60/mo
RAG assistant

8K in / 600 out × 100K requests

$294.42/mo$1280.00/mo
Coding agent

40K in / 4K out × 20K requests

$237.67/mo$1128.00/mo
Document extraction

20K in / 1.5K out × 50K requests

$587.52/mo$2365.00/mo
Bulk classification

500 in / 20 out × 5M requests

$1192.77/mo$4750.00/mo
Run these two through the cost calculator

Which should you pick?

You are running this at volume

DeepSeek V4 Pro 0423

Lowest measured cost per point of capability at $0.0261 per point — the gap compounds with every request.

Quality matters more than the bill

SpaceXAI: Grok 4.5

Highest overall LiveBench score of the two at 75.8.

The workload is coding or agentic work

SpaceXAI: Grok 4.5

Leads on agentic coding — 56.5 against 42.6.

You need to fit large documents in one call

DeepSeek V4 Pro 0423

Wider context window — 1.0M against 500K.

DeepSeek V4 Pro 0423 vs SpaceXAI: Grok 4.5 FAQ

Which is better, DeepSeek V4 Pro 0423 or SpaceXAI: Grok 4.5?

SpaceXAI: Grok 4.5 scores higher, DeepSeek V4 Pro 0423 costs less — it depends on your workload. SpaceXAI: Grok 4.5 is ahead by 4.2 points overall, and DeepSeek V4 Pro 0423 lists 4.5× cheaper per blended million tokens. Whether 4.2 points is worth that depends on how much a wrong answer costs you. DeepSeek V4 Pro 0423 also leads on measured cost per point of capability, at $0.0261 per point.

Is DeepSeek V4 Pro 0423 cheaper than SpaceXAI: Grok 4.5?

DeepSeek V4 Pro 0423 is cheaper. On a 3:1 input:output blend, DeepSeek V4 Pro 0423 lists at $0.665 per million tokens and SpaceXAI: Grok 4.5 at $3.00 — DeepSeek V4 Pro 0423 is 4.5× cheaper. Input and output are priced separately — DeepSeek V4 Pro 0423 charges $0.532 in and $1.06 out, SpaceXAI: Grok 4.5 charges $2.00 and $6.00 — so the model that looks cheaper flips depending on how output-heavy your workload is.

DeepSeek V4 Pro 0423 vs SpaceXAI: Grok 4.5: which scores higher on benchmarks?

DeepSeek V4 Pro 0423 scores 71.6 and SpaceXAI: Grok 4.5 scores 75.8 overall on LiveBench, the mean of its seven categories. That is a 4.2-point lead for SpaceXAI: Grok 4.5. Category scores differ from the overall figure — a model can lead on reasoning and trail on coding, which the per-category table above breaks out.

Which gives better value for money, DeepSeek V4 Pro 0423 or SpaceXAI: Grok 4.5?

DeepSeek V4 Pro 0423. Cost per point divides the measured dollars LiveBench spent running the benchmark by the score it earned, so it captures something token pricing misses: a reasoning model can emit many times more tokens than its per-token rate suggests. DeepSeek V4 Pro 0423 works out at $0.0261 per point and SpaceXAI: Grok 4.5 at $0.0720.

Does DeepSeek V4 Pro 0423 or SpaceXAI: Grok 4.5 have a bigger context window?

DeepSeek V4 Pro 0423 has the larger context window: 1.0M for DeepSeek V4 Pro 0423 against 500K for SpaceXAI: Grok 4.5. 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 DeepSeek V4 Pro 0423 and SpaceXAI: Grok 4.5 support prompt caching?

Both publish a cached-input rate: $0.044 per million for DeepSeek V4 Pro 0423 and $0.300 for SpaceXAI: Grok 4.5, against full input rates of $0.532 and $2.00. 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.
  • ScoresLiveBench 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.