Mistral API Cost Calculator
How much does the Mistral API cost per month?
For Ministral 8B at 200K requests/month, 2,400 input tokens and 350 output tokens per request costs about $40.88 per month after 30% cache use and 100% batch share. Across Mistral's 5 priced models, the cheapest default ranking is Ministral 8B at $81.77 per month.
Pricing data as of June 2026. Sources: Mistral pricing and model documentation. Parameters are shareable in this URL.
Ministral 8B: estimated monthly cost
Formula: provider-specific cached input multiplier × cache rate, plus uncached input and verbosity-adjusted output; cache writes are amortised over the provider TTL, then batch savings are applied.
| Per request | $0.000 |
|---|---|
| Per day | $1.36 |
| Per month | $40.88 |
| Per year | $491 |
OpenAI scenario sensitivity
The default is a server-rendered estimate. Change the cache and batch shares to see when a cheaper qualified model overtakes the selected model; the URL is shareable and preserves the inputs.
| Default | $40.88/month |
|---|---|
| No caching or batch | $81.77/month |
| 50% cache, 50% batch | $61.32/month |
| 100% cache, 100% batch | $40.88/month |
Batch 11 Mistral workload decision depth
1. Mixed synchronous/batch bill
| Async share | Monthly bill | Savings | Eligibility/turnaround |
|---|---|---|---|
| 0% | $82.50 | $0.0000 | Synchronous |
| 25% | $72.19 | $10.31 | Submitted eligible job; asynchronous deadline |
| 50% | $61.88 | $20.63 | Submitted eligible job; asynchronous deadline |
| 100% | $41.25 | $41.25 | Submitted eligible job; asynchronous deadline |
2. Context and modality fit
| Context input | Model | Token bill | Fit/modality boundary |
|---|---|---|---|
| 32K | Ministral 8B | $0.0049 | Window and non-text units: sourced per model or Unavailable |
| 32K | Mistral Small 3.1 | $0.0050 | Window and non-text units: sourced per model or Unavailable |
| 32K | Codestral | $0.0099 | Window and non-text units: sourced per model or Unavailable |
| 32K | Mistral Large 3 | $0.02 | Window and non-text units: sourced per model or Unavailable |
| 128K | Ministral 8B | $0.02 | Window and non-text units: sourced per model or Unavailable |
| 128K | Mistral Small 3.1 | $0.02 | Window and non-text units: sourced per model or Unavailable |
| 128K | Codestral | $0.04 | Window and non-text units: sourced per model or Unavailable |
| 128K | Mistral Large 3 | $0.06 | Window and non-text units: sourced per model or Unavailable |
| 256K | Ministral 8B | $0.04 | Window and non-text units: sourced per model or Unavailable |
| 256K | Mistral Small 3.1 | $0.04 | Window and non-text units: sourced per model or Unavailable |
| 256K | Codestral | $0.08 | Window and non-text units: sourced per model or Unavailable |
| 256K | Mistral Large 3 | $0.13 | Window and non-text units: sourced per model or Unavailable |
3. Large/Medium/Small accepted-result crossover
| Tier | API cost/call | Required uplift | Retry input | Accepted formula |
|---|---|---|---|---|
| Ministral 8B | $0.0004 | User-supplied quality uplift | User-supplied retry rate | API cost ÷ (1 − retry rate) |
| Mistral Small 3.1 | $0.0006 | User-supplied quality uplift | User-supplied retry rate | API cost ÷ (1 − retry rate) |
| Codestral | $0.0010 | User-supplied quality uplift | User-supplied retry rate | API cost ÷ (1 − retry rate) |
Provenance: Batch 11 Mistral delivery/context/accepted-result module; selected model Ministral 8B; inputs are 2,400 input tokens, 350 output tokens, 200,000 calls/month, cache rate 30%, batch share 100%. Verified 2026-06-21. Provider source: https://mistral.ai/pricing · Run this scenario →
Batch 59 · server-rendered evidence boards · verified 2026-09-07
Intent answer: Mistral API billing spans cost-efficient edge models (Ministral 8B) through enterprise flagships (Mistral Large). Batch processing discounts, prompt caching support, and specialized Codestral pricing dictate optimum architecture. Verified 2026-09-07.
Demand evidence: Qualitative demand: Mistral pricing and token calculators reviewed 2026-09-07; exact US monthly volume is unavailable.
Scope boundary: Calculate Mistral API monthly expenses across Mistral Large, Mistral Medium, Mistral Small, Ministral 8B, and Codestral with prompt caching. Exact joins required; unresolved joins render Unavailable.
Mistral tier tiering & workload routing matrix
Deterministic formula / rule: tier_cost = min(input * rate_model + output * rate_model) subject to task complexity threshold.
Boundary: Owns Mistral portfolio routing economics.
| Frozen scenario / field ID | Exact identity and evidence fields | Result | State |
|---|---|---|---|
batch59-calc-mistral-m1-r1simple classification (Ministral 8B) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$simple classification (Ministral 8B); task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 1 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m1-r2general reasoning (Mistral Small) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$general reasoning (Mistral Small); task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 1 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m1-r3complex reasoning (Mistral Large) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$complex reasoning (Mistral Large); task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 1 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m1-r4code completion (Codestral) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$code completion (Codestral); task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 1 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m1-r5multilingual document translation | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$multilingual document translation; task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 1 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m1-r6unsupported model alias | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$unsupported model alias; task token shape; candidate models; unit token costs; monthly bill per model; tier cost delta; selection threshold; verified=2026-09-07 | Unavailable — exact calc-mistral evidence join is not closed for "unsupported model alias" | FAIL CLOSED — manual, probe, or source evidence required |
First-party citation: Mistral AI official pricing. Verified 2026-09-07; missing or conflicting joins fail closed.
Codestral developer team cost projection
Deterministic formula / rule: team_cost = developer_count * daily_requests * (avg_input * rate_in + avg_output * rate_out) * 22_workdays.
Boundary: Owns developer coding assistant budget forecasting.
| Frozen scenario / field ID | Exact identity and evidence fields | Result | State |
|---|---|---|---|
batch59-calc-mistral-m2-r1small team (5 devs, 500 req/day) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$small team (5 devs, 500 req/day); dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 2 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m2-r2medium team (25 devs, 2.5K req/day) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$medium team (25 devs, 2.5K req/day); dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 2 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m2-r3large team (100 devs, 10K req/day) | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$large team (100 devs, 10K req/day); dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 2 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m2-r4automated CI code reviewer | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$automated CI code reviewer; dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 2 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m2-r5batch repository analysis | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$batch repository analysis; dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 2 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m2-r6unsupported IDE plugin rate | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$unsupported IDE plugin rate; dev count; daily queries; prompt context size; completion size; monthly API total; per-seat equivalent cost; verified=2026-09-07 | Unavailable — exact calc-mistral evidence join is not closed for "unsupported IDE plugin rate" | FAIL CLOSED — manual, probe, or source evidence required |
First-party citation: Mistral AI official pricing. Verified 2026-09-07; missing or conflicting joins fail closed.
Mistral prompt caching & batch savings board
Deterministic formula / rule: net_spend = (uncached_inputs * in_rate) + (cached_inputs * cache_rate) + (outputs * out_rate) - batch_discount.
Boundary: Owns Mistral optimization lever modeling.
| Frozen scenario / field ID | Exact identity and evidence fields | Result | State |
|---|---|---|---|
batch59-calc-mistral-m3-r150% cache hit enterprise workload | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$50% cache hit enterprise workload; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 3 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m3-r280% cache hit RAG system | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$80% cache hit RAG system; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 3 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m3-r3asynchronous batch extraction pipeline | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$asynchronous batch extraction pipeline; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 3 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m3-r4streaming interactive zero-cache chat | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$streaming interactive zero-cache chat; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 3 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m3-r5mixed latency-tolerant job queue | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$mixed latency-tolerant job queue; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — frozen calc-mistral fixture requires an exact source, identity, and result receipt | UNTESTED — module 3 rule is reproducible but no production observation is claimed |
batch59-calc-mistral-m3-r6unsupported cache TTL parameter | route=$/llm-cost-calculator/mistral; owner=$calc-mistral; scenario=$unsupported cache TTL parameter; monthly requests; cache hit ratio; standard cost; optimized cost; absolute monthly savings; ROI horizon; verified=2026-09-07 | Unavailable — exact calc-mistral evidence join is not closed for "unsupported cache TTL parameter" | FAIL CLOSED — manual, probe, or source evidence required |
First-party citation: Mistral AI official pricing. Verified 2026-09-07; missing or conflicting joins fail closed.
Method and limitations: this board exposes deterministic rules, first-party citations, and dated evidence identities. It does not invent volume, coverage, entitlement, retention, residency, quota, capacity, feature support, quality, price, reliability, or legal conclusions. Run the calc-mistral evidence flow →
Ranked cost — 200K requests/month
| Model | Provider | List monthly | Effective monthly | Verbosity | Rank Δ |
|---|---|---|---|---|---|
| Ministral 8Bbudget | Mistral | $82.50 | $40.88 | 0.93× | — |
| Mistral Small 3.1budget | Mistral | $114 | $53.85 | 0.85× | — |
| Codestralbudget | Mistral | $207 | $96.89 | 0.79× | — |
| Mistral Large 3budget | Mistral | $345 | $173 | — | — |
| Mistral Medium 3mid | Mistral | $1,245 | $580 | 0.84× | — |
Levers live on Mistral
The billing gotcha
Mistral's batch discount applies per submitted job, not per call, so it only pays off once you're already batching requests server-side rather than calling the synchronous API in a loop — turning batch on in the estimator without changing your actual call pattern will overstate your real saving, since a request sent one at a time never qualifies for the batch rate no matter what this page's toggle shows.