Quick Answer
TeamoRouter's deepseek-v4-pro-free (and deepseek-v4-flash-free) provides 200 requests per account per day, reset daily, over an OpenAI-compatible endpoint.
The key point: the quota counts requests, not tokens. That's why it goes further than it sounds. One "request" is one full model call — the model can read 100K+ tokens, reason, and produce a long diff, and that's one unit against the 200.
Requests vs Tokens: A Critical Distinction
Most APIs bill by token (input + output). The free tier bills by request:
| Billing unit | How it's counted | Trait |
|---|---|---|
| Token | Input + output token count | Longer output = more expensive |
| Request (free tier) | One full turn | Independent of output length |
So "read 100K context + output a 5,000-token diff" and a one-word "hello" both count as 1 request. That's hugely favorable for agent work, where a single request packs in lots of context and produces long output.
What One "Request" Means Inside an Agent Loop
In plain chat, one request = one Q&A. In an agent, the model calls tools, and each tool round-trip is its own request:
Model: "read src/reporting.py" ← request 1
Tool: returns file contents
Model: "edit line 45, dedupe branches" ← request 2
Tool: write done
Model: "run tests" ← request 3
...
A multi-step refactor can burn 10–15. So 200 requests really means 200 rounds of model decision-making, not 200 tokens. Comfortable for a solo developer's day, but letting the agent thrash drains it fast.
What 200 Requests Covers
A week of normal use:
| Task | Approx. requests |
|---|---|
| Read a file + explain a function | 1–3 |
| Single-file edit (read + change + verify) | 3–5 |
| Multi-file refactor (cross-file + run tests) | 10–15 |
| Long agent loop (build + self-test + fix) | 20–40 |
Bottom line: single-file daily edits are fully covered (a dozen+), heavy agent loops cover 3–5 hard tasks. Past that, paid is cheap.
When the Quota Runs Out: Switch to Paid
Once a free tier hits zero, swap the model ID — paid models have no account-level limit:
# from free to paid
deepseek-v4-pro-free → deepseek-v4-pro
deepseek-v4-flash-free → deepseek-v4-flash
Paid pricing: deepseek-v4-pro $1.74/$3.48 (input/output per 1M tokens), deepseek-v4-flash $0.14/$0.28 — flat, no peak windows. Even heavy use stays under a few dollars a day.
FAQ
Does the quota reset daily or accumulate?
Resets daily. Per-account, not per-session; multiple sessions share one 200-request pool.
Is the free tier the same capability as paid?
Yes — the difference is rate limits. -free caps at 200/day; paid has no account limit. deepseek-v4-pro currently resolves to DeepSeek-V4-Pro-0813.
Does longer output cost more quota?
No. Request-based means one request regardless of output length — which rewards giving full context and asking for complete results, not splitting into many small calls.
Does it work for teams?
The free quota is per account, no team pooling. Each member gets their own 200. For team throughput, go paid.
How do I check what's left?
See the TeamoRouter rate-limits docs. At the limit, requests return a rate-limit error — switch to paid.
Register at TeamoRouter and take 200 free V4 Pro requests daily, then pay flat past that.