DeepSeek Guide — whale logoDeepSeek GuideFAN SITE
MODEL GUIDE4 MIN READ

DeepSeek V4 Pro 0813 Build Explained: What Changed at GA

UPDATED: AUG 16, 2026AUTHOR: INDEPENDENT FAN GUIDE
OVERVIEW

V4 Pro 0813 is the GA checkpoint of Aug 13, 2026: SWE-bench 96.4%, TB2.1 87.9, native Responses API, reasoning effort levels, and new pricing.

01

What '0813' Means

`deepseek-v4-pro-0813` is the checkpoint DeepSeek shipped on August 13, 2026 — the build that moved V4 Pro from preview to generally available. '0813' is the release date in YYMMDD-style shorthand, matching the Flash-0731 naming convention from July 31[1][4].

The API keeps the stable alias `deepseek-v4-pro`, which now resolves to the 0813 weights. On OpenRouter the GA build is listed separately as `deepseek/deepseek-v4-pro-0813` (created August 13) alongside the older preview entry `deepseek/deepseek-v4-pro`, so if you use a third-party router, check which entry your config points at[10].

The July 31 Flash update established the pattern: same architecture, retrained weights, new checkpoint name. The 0813 Pro build follows the same logic — a materially improved checkpoint, not a config tweak.

  • 0813 = GA checkpoint released Aug 13, 2026
  • API alias deepseek-v4-pro resolves to 0813 weights
  • OpenRouter lists it as deepseek/deepseek-v4-pro-0813
  • Same architecture as preview: 1.6T total / 49B active MoE
NOTE

Do not confuse the 0813 GA checkpoint with the preview-era entries still listed on some routers and pricing pages — check the created date[10].

02

What Changed at GA

The GA release bundled three headline changes plus a pricing overhaul[1].

The reasoning-effort and Responses API changes are the ones developers actually feel day-to-day. `reasoning_effort` accepts `low` / `high` / `max` and defaults to `high`; the Responses API works with Codex clients through DeepSeek's one-click setup script[4][5].

Pricing is the other big shift: peak/off-peak rates took effect 16:00 UTC on August 16, three days after GA. Every V4 Pro line costs more than the old flat rate, with output at $1.98 off-peak / $3.96 peak[2].

ChangePreview era0813 GA
Agent capabilityPreview-level agent scoresMajor upgrades, 'strong production gains'
Reasoning effortFixed effort behaviorlow / high / max, default high
OpenAI Responses APIFlash onlyNative support, Codex one-click setup
App / web accessModel menu onlyV4 Pro via 'Expert Mode'
PricingFlat $0.435 / $0.87Peak & off-peak rates from Aug 16
NOTE

Model names and endpoint URLs are unchanged — existing preview-era integrations keep working against the 0813 weights[1][3].

03

0813 on the Benchmarks

The model card reports SWE-bench Verified at 96.4% — the headline coding number — alongside the full agent table: Terminal Bench 2.1 at 87.9, Cybergym 83.3 (first), AutomationBench 31.8 (first)[4].

Note the SWE-bench number: the preview-era card reported 80.6% (as of August 1), and the 0813 card reports 96.4% — a 15.8-point jump that matches the pattern across the rest of the table[4].

Independent testing (MindStudio, 8 tasks) scored the 0813 build at 76.25%, versus 24.8% for the preview — same direction, more conservative magnitude[6]. The full benchmark page has the complete tables and the harness caveat.

Benchmark0813 GAPro PreviewDelta
SWE-bench Verified96.4%80.6%+15.8 pts
Terminal Bench 2.187.972.1+15.8
Cybergym83.352.7+30.6
DeepSWE62.712.8+49.9
Toolathlon-Verified74.155.9+18.2
04

Model Card & Self-Hosting Notes

The 0813 model card on Hugging Face documents the weights, the encoding folder, and the official vLLM deployment recipe — including the DSpark speculative-decoding flag[4].

Self-hosting 0813 is a serious lift — third-party estimates put the FP4+FP8 weights at roughly 862GB of VRAM. The hosted API is the practical route for most teams; the model size page has the full hardware math.

  • No Jinja chat template on the card — encoding is provided as a Python folder demonstrating OpenAI-compatible format encoding
  • vLLM deployment: enable DSpark with a single flag: --speculative-config '{"method":"dspark",...}'
  • Reference hardware: 4xGB300 single node for the full model
  • Weights are MIT-licensed; self-hosting is officially supported
NOTE

The card's internal DSBench rows are marked with a dagger — treat them as vendor-internal numbers[4].

05

Should You Upgrade to 0813?

If you are on the preview, the upgrade is automatic — the stable model ID already points at 0813. The real decision is budget and workload[2][6].

Upgrade if your work is agentic or coding-heavy: the GA build's agent gains are the largest single-version jump in the family's history, and the Responses API unlocks Codex workflows that preview-era Pro could not do[1][6].

Hold the line if your workload is high-volume and latency-sensitive chat: independent tests note 0813 over-thinks simple problems, and for everyday tasks V4 Flash often feels snappier at one-third the price[6]. The peak-hour rates make that calculus sharper — see the pricing page.

NOTE

The 0813 checkpoint is the current GA as of August 16, 2026. DeepSeek has not announced the next build.

Sponsored
Sponsored