DeepSeek desk
A date-ordered index of DeepSeek coverage across releases, API changes, pricing, local deployment and the surrounding developer ecosystem. This archive now includes 34 articles, including a migrated editorial collection from our sister site, deepseekv4pro.com. Each entry preserves its original publication link and primary source.
Official
DeepSeek has published a new V4 API rate card that takes effect at 00:00 Beijing time on August 17. The change introduces peak and off-peak pricing for V4 Flash and V4 Pro, making model routing, caching, and job timing more consequential for API teams.
Official
DeepSeek Harness is now available in developer preview. The open-source agent framework ships as dsh, starts from npm, and gives developers a plugin-based route to model configuration, workspace tools, delegated tasks, and Web UI workflows.
Official
DeepSeek-V4-Pro-0813 is now the build behind deepseek-v4-pro. It keeps the familiar API and pricing, but changes the question developers should ask: how much work can this route now take before it needs a more expensive fallback?
Official
DeepSeek V4 Flash 0731 is now in public beta with stronger agent-focused performance claims. Independent analysis also places the release on a more competitive intelligence-versus-cost curve.
Official
Checked on July 14, 2026: DeepSeek's official docs sitemap currently lists dedicated chat_curl, chat_nodejs, and chat_python sample pages, and all three use the current https://api.deepseek.com endpoint plus deepseek-v4-pro instead of older alias-first snippets.
Official
Checked on July 13, 2026: DeepSeek's official docs sitemap currently lists both a top-level Prompt Library page and an Integrate with AI Tools guide covering Claude Code, OpenCode, and OpenClaw, giving developers a first-party docs path from prompt exploration to concrete tool configuration.
Architecture
Flash lists at $0.14 per million input tokens because it does far less arithmetic per token than its 284B size suggests. A walk through the three design choices — sparse experts, latent attention, and disk-backed caching — that put that number on the page.
Agents
Agents are the workload that turns a cheap model into an expensive invoice, because the loop re-sends everything on every turn. The token math of an agent loop, the caching that fixes it, and the routing and retry policy that keep a Flash agent honest.
Official
Checked on July 8, 2026: DeepSeek's official docs sitemap currently lists dedicated thinking-mode sample pages for both non-streaming and streaming chat completions, clarifying when reasoning_content is just captured, when it is ignored in ordinary follow-up turns, and how delta.reasoning_content should be accumulated separately from visible answer text.
Official
Checked on July 6, 2026: DeepSeek's official docs currently expose both a Python thinking-mode tool-call sample and a separate output transcript page, giving developers a first-party reference for tool schemas, chained calls, blank content turns, and reasoning_content replay behavior.
Official
Checked on July 5, 2026: DeepSeek's official homepage still anchors the official @deepseek_ai X presence, but the strongest current first-party developer signal is in the API docs and changelog, which both keep the July 24, 2026 retirement warning for deepseek-chat and deepseek-reasoner while pointing users to explicit deepseek-v4-flash and deepseek-v4-pro model names.
API
DeepSeek's hosted API drops the legacy deepseek-chat and deepseek-reasoner aliases on 2026-07-24 at 15:59 UTC. After that, both IDs route to deepseek-v4-flash.
Ecosystem
The official Ollama library now lists deepseek-v4-flash, and OpenClaw users can wire it up with a single launch command.
Recap
Provider prices settled near $0.09–0.14 input, the July 24 alias retirement was confirmed, and the tooling ecosystem caught up.
Official
Checked on July 2, 2026: OpenAI's Codex docs now document custom model providers through model_provider, model_providers, base_url, and env_key. That means DeepSeek can be framed as a third-party Codex route: use our DeepSeek access, point Codex at the configured DeepSeek provider, and keep the known Responses API, profile, Desktop picker, and V4 thinking-mode checks in place.
Official
Checked on July 2, 2026: DeepSeek's official API reference documents a GET /models endpoint that currently examples only deepseek-v4-flash and deepseek-v4-pro, plus a GET /user/balance endpoint with is_available, currency, granted balance, and topped-up balance fields for direct budget checks.
Official
Checked on June 30, 2026: DeepSeek's official Token & Token Usage page says tokens are the billing unit, gives rough heuristics of about 0.3 token per English character and 0.6 token per Chinese character, and links a deepseek_tokenizer.zip package for offline token estimation while still treating API usage output as the real source of truth.
Official
Checked on June 29, 2026: DeepSeek's official Tool Calls guide documents OpenAI-compatible function calling, a Beta-only strict mode for exact schema matching, concrete supported and unsupported JSON Schema rules, and a required tool role replay pattern after execution.
Official
Checked on June 28, 2026: DeepSeek's official Models & Pricing and Rate Limit pages currently show a 1M context window, 384K max output, per-1M token pricing, and account-level concurrency limits of 2500 for DeepSeek V4 Flash and 500 for DeepSeek V4 Pro, with 429s when that cap is exceeded.
Official
Checked on June 27, 2026: DeepSeek's official multi-round conversation guide says /chat/completions does not store context server-side, so each follow-up call must resend the prior conversation history, including earlier assistant output, before the next user turn.
Official
Checked on June 26, 2026: DeepSeek's official WorkBuddy or CodeBuddy page now documents exact user-level and project-level models.json paths, https://api.deepseek.com/v1/chat/completions as the endpoint, a UTF-8 without BOM requirement, and a restart-first troubleshooting flow for V4 Pro and V4 Flash.
Official
Checked on June 25, 2026: DeepSeek's official Anthropic API page says claude-opus* maps to V4 Pro, claude-haiku* and claude-sonnet* map to V4 Flash, unsupported model names automatically fall back to Flash, and the mapping can help bypass Claude Desktop developer-mode model-name restrictions.
Official
Checked on June 24, 2026: DeepSeek's official GitHub Copilot CLI page says teams should use the Anthropic-compatible endpoint, warns that the OpenAI provider triggers a 400 reasoning-content error, sets explicit token ceilings for deepseek-v4-pro, and clarifies that offline mode still sends prompts to DeepSeek.
Official
Checked on June 23, 2026: DeepSeek's official Claude Code page now documents the Anthropic-compatible environment contract, native Web Search support, and an explicit mapping where Claude-style model names route to DeepSeek V4 Pro or Flash.
Official
Checked on June 22, 2026: DeepSeek's official OpenClaw page now documents openclaw onboard --install-daemon, QuickStart provider selection, manual entry of deepseek-v4-pro or deepseek-v4-flash, and follow-up launch routes through the dashboard, TUI, or terminal.
Official
Checked on June 20, 2026: DeepSeek's official Reasonix integration page now documents npx reasonix code, stores the API key in ~/.reasonix/config.json, defaults to DeepSeek V4 Flash, and exposes /pro plus /preset max for stronger reasoning when needed.
Official
Checked on June 18, 2026: DeepSeek's official research and GitHub surfaces already include multimodal work through Janus-Pro and DeepSeek-VL2, covering image understanding, OCR, document/chart reasoning, visual grounding, and text-to-image generation. Official V4 API docs still list V4 Pro and V4 Flash as text-first API routes, and DeepSeek's Copilot page says V4 itself is text-only with an optional vision proxy, so this should be framed as DeepSeek's multimodal ecosystem expanding rather than a newly confirmed native V4 Vision API.
Official
Checked on June 18, 2026: DeepSeek's official Pi integration guide now gives a concrete models.json provider block with api.deepseek.com, 1M context metadata, Pro and Flash pricing fields, and a reasoning-effort map that translates Pi's levels to DeepSeek's supported behavior.
Official
Checked on June 11, 2026: DeepSeek's official docs now give a cleaner production picture for V4 traffic management, with 500 Pro concurrency, 2500 Flash concurrency, explicit user_id isolation rules, and request keep-alive handling.
Official
DeepSeek's current pricing page now says V4-Pro API pricing will be officially adjusted to one quarter of the original price after the 75% discount ends on May 31, 2026. Treat this as an official price reset, not a guarantee that prices can never change.
Official
DeepSeek's official docs now tie Claude-style integrations more clearly to exact V4 model strings, while the main API quick-start warns that deepseek-chat and deepseek-reasoner will be deprecated on July 24, 2026.
Pricing
The official page separates cache-hit, cache-miss input, and output pricing for V4 Flash and V4 Pro, and warns that older aliases will be deprecated.
Release
The model card describes V4 as a preview release: Pro at 1.6T total / 49B active, Flash at 284B total / 13B active, both with 1M context.
Correction
DeepSeek V4 Pro is listed at $1.74/M input and $3.48/M output. The $12/M figure circulating in some comparisons is Gemini 3.1 Pro Preview's output price.