// head_to_head

Mistral Medium 3.5 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

Mistral Medium 3.5 and SpaceXAI: Grok 4.5 are priced within ~10% of each other.

Mistral Medium 3.5 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.

mistralai

Mistral Medium 3.5

Blended / 1M
$3.00
Context
262K
Released
Apr 30, 2026
Overall score
Not evaluated
reasoningtool callingimage inputfile input

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

MetricMistral Medium 3.5SpaceXAI: Grok 4.5
LiveBench overall

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

75.8
Cost per point

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

$0.0720
Blended price / 1M

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

$3.00$3.00
Input price / 1M$1.50win$2.00
Output price / 1M$7.50$6.00win
Cached input / 1M

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

$0.300
Context window262K500Kwin
Max output tokens210K450Kwin

Benchmarks by category

An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — Mistral Medium 3.5 on top, SpaceXAI: Grok 4.5 below, both out of 100.

Agentic coding
56.5
Coding
68.6
Reasoning
87.2
Mathematics
90.8
Data analysis
73.0
Language
82.8
Instruction following
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.

WorkloadMistral Medium 3.5SpaceXAI: Grok 4.5
Support chatbot

1.2K in / 400 out × 200K requests

$960.00/mo$837.60/mo
RAG assistant

8K in / 600 out × 100K requests

$1,650/mo$1,280/mo
Coding agent

40K in / 4K out × 20K requests

$1,800/mo$1,128/mo
Document extraction

20K in / 1.5K out × 50K requests

$2,063/mo$2,365/mo
Bulk classification

500 in / 20 out × 5M requests

$4,500/mo$4,750/mo
Run these two through the cost calculator

Which should you pick?

You need to fit large documents in one call

SpaceXAI: Grok 4.5

Wider context window — 500K against 262K.

Mistral Medium 3.5 vs SpaceXAI: Grok 4.5 FAQ

Which is better, Mistral Medium 3.5 or SpaceXAI: Grok 4.5?

Mistral Medium 3.5 and SpaceXAI: Grok 4.5 are priced within ~10% of each other. Mistral Medium 3.5 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 Mistral Medium 3.5 cheaper than SpaceXAI: Grok 4.5?

They cost about the same. Both land near $3.00 per million tokens on a 3:1 input:output blend, so price is unlikely to be the deciding factor between them.

Does Mistral Medium 3.5 or SpaceXAI: Grok 4.5 have a bigger context window?

SpaceXAI: Grok 4.5 has the larger context window: 262K for Mistral Medium 3.5 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 Mistral Medium 3.5 and SpaceXAI: Grok 4.5 support prompt caching?

SpaceXAI: Grok 4.5 publishes a cached-input rate of $0.300 per million tokens against a full input rate of $2.00. The catalogue lists no separate cached rate for Mistral Medium 3.5, which means the provider does not price it separately here — not that caching is unavailable.

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.