European mid tier versus Google's volume Flash at matching input prices.
mistral-medium-latest · mistral
gemini-3.5-flash · google
Benchmark
| Workload Scenario | Mistral Medium 3.5 | Gemini 3.5 Flash | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $1.50 | $1.50 | — |
| 1M output tokens (generation) | $7.50 | $9.00 | −$1.50 |
| 1M tokens · 70% input / 30% output mix | $3.30 | $3.75 | −$0.45 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.0093 | $0.0105 | −$0.0012 |
| Monthly scale (10K requests / day) | $2,790.00 | $3,150.00 | −$360.00 |
Negative difference = Mistral Medium 3.5 is cheaper. Positive = Gemini 3.5 Flash is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
Gemini 3.5 Flash has the cheaper input rate, while Mistral Medium 3.5 has the cheaper output rate. The crossover happens when output makes up about 0% of your total tokens.
Below that share (retrieval, summarization, extraction — lots of context in, little text out) Gemini 3.5 Flash is cheaper. Above it (generation, translation, coding — long completions) Mistral Medium 3.5 wins.
FAQ
On input, Gemini 3.5 Flash is cheaper ($1.50/M vs $1.50/M). On output, Mistral Medium 3.5 is cheaper ($7.50/M vs $9.00/M). For workloads where more than 0% of tokens are output, the output-cheaper model wins overall.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.0093 on Mistral Medium 3.5 and $0.0105 on Gemini 3.5 Flash — Mistral Medium 3.5 is 1.1× cheaper for that workload.
Mistral Medium 3.5 supports 131,072 tokens (32.8K max output); Gemini 3.5 Flash supports 1,048,576 (65.5K max output). Gemini 3.5 Flash fits 8.0× more context, which matters for long documents and agents.