Mistral Small 4
by Mistral · mistralai/mistral-small-2603 · #75 cheapest of 325 paid models
Prices updated Jul 22, 2026, 2:19 AM UTC · refreshed hourly
Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system. It combines strong reasoning from...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.15
per 1M tokens
Output
$0.60
per 1M tokens
Blended (3:1)
$0.262
3 input : 1 output
Cache read
$0.015
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Mistral Small 4 with prompt caching.
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.262 | $0.0159 | — |
| 50% | $0.212 | $0.00915 | −42% |
| 90% | $0.171 | $0.00375 | −76% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $0.000255 | $0.255 |
| Document summary | 8,000 | 1,000 | $0.0018 | $1.80 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.0057 | $5.70 |
| Long-context analysis | 150,000 | 5,000 | $0.0255 | $25.50 |
Performance
Intelligence Index
19.6
AA composite quality
Coding Index
26.6
Math Index
—
Agentic Index
4.7
Output speed
175 tok/s
median
Time to first token
0.53s
median
Time to first answer
11.93s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 76.9% |
| Humanity's Last Exam | 9.5% |
| SciCode | 38.0% |
| IFBench | 48.2% |
| AA-LCR (Long Context Reasoning) | 44.7% |
| Terminal-Bench Hard | 17.4% |
| Terminal-Bench 2.1 | 21.0% |
| τ²-Bench (Telecom) | 41.2% |
| τ³-Bench Banking | 5.2% |
Specs
Context window
262K
tokens
Max output
—
tokens
Input modalities
text, image
Tokenizer
Mistral