DeepSeek V4 Flash as a free coding agent in your terminal

DeepSeek V4 Flash is the model a new Clixad account starts on, and one of five that are free every day. That is about 49 turns a day of agent work in your terminal, with no DeepSeek account, no API key and no card.

  • No API key
  • No monthly subscription
  • Free models every day
npm install -g clixad

Node 20 or newer. Then clixad login signs you in with GitHub, and clixad in a project opens the agent.

49free turns a day
484credits a turn past that
1,048,576tokens of context
$0to start

DeepSeek V4 Flash on Clixad, in numbers

A turn is 20,000 input and 1,500 output tokens, the size the /model picker estimates with. Recomputed from packages/pricing by a test.
ModelContext window, tokensCredits per turnFree turns a dayTurns on the signup bonus
DeepSeek V4 Flash1,048,5764844951

The free turns come out of the daily allowance of 24,000 credits, which is shared with the other four free models and sits inside a daily budget shared by everybody, so on a busy day they can end sooner. After that a turn costs 484 credits until 00:00 UTC. The 25,000 credits a new account starts with cover 51 more, and further credits come from completing an offer in a browser tab, credited on completion, or from a one-time pack. A screenout, where a survey ends part way with nothing or a small bonus, is the most common outcome of starting a survey.

V4 Flash, not V4.1 Flash

DeepSeek's own API has moved on. Its pricing page lists deepseek-flash as DeepSeek-V4.1-Flash, and says of the old name: "The legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired, their requests are served by the DeepSeek-V4.1-Flash model and billed at the Flash price."

Clixad's DeepSeek V4 Flash is the earlier model, served through OpenRouter by providers that still host it. Clixad has not moved to V4.1 Flash. If you want DeepSeek's newest Flash model, DeepSeek's own API serves it.

With your own DeepSeek key instead

DeepSeek sells its API per token, from a balance you top up. For V4.1 Flash its pricing page lists $0.15 per million input tokens on a cache miss and $0.60 per million output tokens off-peak, and twice that in peak hours. Aider, Cline and OpenCode can all use a DeepSeek key.

If you already hold a DeepSeek balance, that is the cheaper way to do a lot of work, since credits bought from us carry a 1.5 markup over provider cost. Clixad is for starting without any of that: no account with DeepSeek, no top-up, and 49 free turns a day before any money is involved. It also puts Claude, GPT, Gemini and Kimi models on the same balance for the task that needs a stronger one.

When V4 Flash is not enough

Switch with /model in the middle of a session. Kimi K2.5 is also free, at 2,352 credits a turn, which is about 10 turns a day from the same allowance. Past the free set, Claude Sonnet 5.5 and GPT-6 Sol cost 8,250 credits a turn and Claude Opus 5.5 costs 16,500, and while a model call runs the terminal can show a line marked Ad. Nothing you type takes part in choosing it, and CLIXAD_SPONSOR=0 switches it off.

FAQ

Can I use DeepSeek V4 Flash for coding for free?

Yes. It is one of five models free in Clixad up to 24,000 credits of usage a day, which is about 49 turns of 20,000 input and 1,500 output tokens. It needs no DeepSeek account and no API key.

Do I need a DeepSeek API key?

No. Clixad's gateway holds the provider keys and routes DeepSeek V4 Flash through OpenRouter. You sign in with GitHub.

Is it DeepSeek V4 Flash or V4.1 Flash?

V4 Flash. DeepSeek's own API now serves V4.1 Flash under the name deepseek-flash and has retired V4 Flash there. Clixad serves V4 Flash through OpenRouter, from providers that still host it.

What does DeepSeek V4 Flash cost after the free turns?

484 credits for a turn of 20,000 input and 1,500 output tokens, until the free allowance resets at 00:00 UTC.

Sources

  1. DeepSeek, Models & Pricing, for deepseek-flash as DeepSeek-V4.1-Flash, the retirement of the legacy name deepseek-v4-flash in the quoted sentence, and the off-peak prices of $0.15 per million input tokens on a cache miss and $0.60 per million output tokens, with "Off-peak rates are half of the peak rates." Read 29 September 2026.
  2. Cline, Pricing, which lists DeepSeek among its providers. Aider, aider.chat, which names DeepSeek models among those it works with. OpenCode, opencode.ai, which connects "any model from any provider". All read 29 September 2026.
  3. The model a new account starts on: DEFAULT_MODEL_ID in packages/pricing. Its context window: contextTokens, read from OpenRouter's model list on 27 September 2026.
  4. Every Clixad figure: packages/pricing (creditsForUsage, signupBonusCredits, freeTierDailyCreditsPerUser, FREE_MODEL_IDS, maxAdRewardUsdPerDay, CREDIT_PACKS), over the model catalogue read from OpenRouter on 27 September 2026, recomputed by the test beside this page. A turn is 20,000 input and 1,500 output tokens, FREE_TURN_USAGE, the size the /model picker estimates with.