Skip to content
AI.info

Model Tracker

DeepSeek-Prover-V1.5

DeepSeek-Prover-V1.5 is an AI model made by DeepSeek AI. It is generally available. Its weights are open.

DeepSeek AI

Released
Not confirmed: The maker's pages provide a paper submission date, but do not state that date as the model's first availability or release date.
Status
Generally available — github.com, read 16 October 2026
Takes
Text, Code — github.com, read 16 October 2026
Returns
Not confirmed: two readings disagree: code and text, code
Open weights
Yes — github.com, read 16 October 2026
Licence
Not confirmed: two readings disagree: DeepSeek License Agreement and MIT (code); DeepSeekMath Model License (model)
Context window
Not confirmed: The maker's pages do not state the standard context-window size.
Longest answer
Not confirmed: The maker's pages do not state a standard maximum output limit; the quick-start example's sampling limit is not identified as the model's standard limit.
Knowledge cutoff
Not confirmed: The maker's pages do not state a knowledge-cutoff date.
API price, input
Not confirmed: The maker's pages do not state a standard API input price for this model.
API price, output
Not confirmed: The maker's pages do not state a standard API output price for this model.
Where
Downloadable weights — github.com, read 16 October 2026
Parameters
7,000,000,000 — github.com, read 16 October 2026
Model card
Not confirmed: The maker links separate cards for the Base, SFT, and RL variants, but does not identify one model card for the unsuffixed DeepSeek-Prover-V1.5 family entry.

Questions

Is DeepSeek-Prover-V1.5 still available?

Yes: it is generally available, as DeepSeek AI states it.

Are the weights of DeepSeek-Prover-V1.5 open?

Yes: DeepSeek AI publishes the weights of DeepSeek-Prover-V1.5 for download.

How many parameters does DeepSeek-Prover-V1.5 have?

DeepSeek-Prover-V1.5 has 7,000,000,000 parameters, as DeepSeek AI publishes it.

More from DeepSeek AI

Sources

  1. DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search (read 16 October 2026)
  2. quick_start.py (read 16 October 2026)
  3. LICENSE-MODEL (read 16 October 2026)

All of the Model Tracker