> Cohere
Cohere's API splits keys into a free Trial tier (auto-created at sign-up: 1,000 API calls/month, 20 RPM on chat models, no commercial use) and an application-gated Production tier (500 RPM chat, pay-as-you-go, invoiced monthly or at $250 outstanding). Legacy Command models price from $0.30–$2.50 per 1M in; dedicated Model Vault instances run $4–$10/hr.
// pros
- A genuinely free trial key, auto-created at sign-up: 1,000 API calls/month (monthly reset) and 20 RPM on chat models — no card, no application for the trial itself.
- Published per-endpoint trial limits — Embed 2,000 inputs/min, Rerank 10/min, Parse 500/min, audio 5/min — so you can plan exactly what fits.
- An affordable legacy line for RAG plumbing: Command-Light $0.30/$0.60, Command $1/$2, Aya Expanse $0.50/$1.50 per 1M in.
- Monthly invoicing with a $250 outstanding-balance threshold — small workloads aren't billed in weekly drips.
- Dedicated Model Vault from $4/hr (Embed 4) to $10/hr (Rerank 4 Pro Large) for guaranteed, non-shared inference.
// cons
- Trial keys are explicitly barred from production or commercial use — the free tier is an evaluation tier, full stop.
- Production keys require an application (Owner role + Go-to-Production workflow) — no self-serve jump from trial to paid scale.
- Current-generation models (Command A+, A Reasoning, A Vision) list no published production rate limit — 'contact sales' is the ceiling.
- The flagship generative line has moved to the North platform (custom enterprise pricing) — the published API tables are legacy-weighted.
Use the trial key as designed: evaluation, integration tests, low-volume internal tooling. For anything customer-facing, start the production application early — and expect a sales conversation for the current-gen models.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
How we estimated this: the free allowance is converted to a dollar value using the provider’s own pay-as-you-go rates for the models a typical developer would reach for. When the allowance is metered in requests, tokens or minutes rather than dollars, we assume median usage of the cheapest capable model. Because assumptions are involved, the figure carries the est. flag — treat it as an order of magnitude, not a quote.
note: Trial key (auto-created at sign-up): 1,000 calls/month + 20 RPM chat (docs.cohere.com rate-limits, verified 2026-09-16); no production/commercial use. Dollar value is an estimate.
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Free tier (source-reported) | free | — | — | — | Trial key allows 1000 API calls/month (resets monthly, no card required); Chat capped at ~20 requests/minute; trial keys can access all models but are not permitted for production or commercial use. |
| Trial Key | free | — | — | — | 1000 calls per month; non-commercial use only; rate-limited. |
| Production (Command R) | payg | — | $0.500 | $1.50 | Production rate for Command R. |
| Production (Command R+) | payg | — | $2.50 | $10.00 | Production rate for Command R+ 08-2024. |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| Command R (08-2024) | $0.150 | $0.600 | 128k |
| Command R+ (08-2024) | $2.50 | $10.00 | 128k |
| Command R+ | — | — | — |
| Command R7B (and all Cohere models — trial and production have the same model access) | — | — | — |
| Command A | — | — | 256k |
$ Does Cohere have a free tier?
Yes — Cohere offers a free tier that we value at roughly $2.00 in free credits (estimate — derived from rate-limited usage at pay-as-you-go prices). Always confirm current limits on Cohere's official pricing page — free tiers change without notice.
$ How much does Cohere cost per 1M tokens?
The cheapest model we track at Cohere is $0.150 per 1M input tokens and $0.600 per 1M output tokens. This is normalized from Cohere's published pricing — verify with the provider before purchasing, since prices change frequently.
$ Is Cohere safe to use?
Uprouter rates Cohere as low risk. Low risk: first-party, with published trial/production rate limits and an explicit free-tier use policy — no commercial use on trial keys is stated, not hidden (verified 2026-09-16: 1,000 calls/month, 20 RPM chat). The caveat is the trial→production gate: current-gen model access sits behind 'contact sales', which caps self-serve scale. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
full history →best-effort probes · not a guarantee · last 15h ago
| Checked | Status | Latency |
|---|---|---|
| 15h ago | up | 65ms |
| 1d ago | up | 82ms |
| 3d ago | up | 58ms |
| 4d ago | up | 54ms |
| 6d ago | up | 94ms |
| 7d ago | up | 199ms |
| 9d ago | up | 56ms |
| 10d ago | up | 51ms |
| 12d ago | up | 47ms |
| 13d ago | up | 57ms |
| 14d ago | up | 56ms |
| 15d ago | up | 49ms |
| 15d ago | up | 50ms |
| 16d ago | up | 53ms |
| 17d ago | up | 62ms |
| 18d ago | up | 76ms |
| 18d ago | up | 51ms |
| 19d ago | up | 47ms |
| 20d ago | up | 131ms |
| 21d ago | up | 52ms |
| 22d ago | up | 49ms |
| 22d ago | up | 44ms |
| 23d ago | up | 52ms |
| 24d ago | up | 48ms |
| 25d ago | up | 67ms |
| 26d ago | up | 84ms |
| 26d ago | up | 89ms |
| 27d ago | up | 56ms |
| 28d ago | up | 61ms |
| 29d ago | up | 64ms |
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low risk: first-party, with published trial/production rate limits and an explicit free-tier use policy — no commercial use on trial keys is stated, not hidden (verified 2026-09-16: 1,000 calls/month, 20 RPM chat). The caveat is the trial→production gate: current-gen model access sits behind 'contact sales', which caps self-serve scale.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# Cohere's own API — get a key from their dashboard
curl https://YOUR-PROVIDER-ENDPOINT/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "command-r",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.