Mistral Medium 3.5
by Mistral · mistralai/mistral-medium-3-5 · #323 cheapest of 422 paid models
Prices updated Sep 19, 2026, 6:21 PM UTC · refreshed hourly
Mistral Medium 3.5 is a dense 128B instruction-following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$1.50
per 1M tokens
Output
$7.50
per 1M tokens
Blended (3:1)
$3.00
3 input : 1 output
Cache read
—
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Mistral Medium 3.5 (no cache-read rate published — hits billed as normal input).
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $3.00 | $0.1605 | — |
| 50% | $3.00 | $0.1605 | −0% |
| 90% | $3.00 | $0.1605 | −0% |
*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 | $0.003 | $3.00 |
| Document summary | 8,000 | 1,000 | $0.0195 | $19.50 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.06 | $60.00 |
| Long-context analysis | 150,000 | 5,000 | $0.2625 | $262.50 |
Performance
Intelligence Index
14.9
AA composite quality
Coding Index
46.9
Math Index
—
Agentic Index
9.4
Output speed
149 tok/s
median
Time to first token
0.66s
median
Time to first answer
14.13s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 74.8% |
| Humanity's Last Exam | 13.8% |
| SciCode | 40.2% |
| IFBench | 68.8% |
| AA-LCR (Long Context Reasoning) | 69.3% |
| Terminal-Bench Hard | 33.3% |
| Terminal-Bench 2.1 | 50.6% |
| τ²-Bench (Telecom) | 94.2% |
| τ³-Bench Banking | 15.1% |
Specs
Context window
262K
tokens
Max output
210K
tokens
Input modalities
text, image, file
Tokenizer
Mistral