# DeepSeek V3 Fast

Model id on AIHubMix: `DeepSeek-V3-Fast`
Create an API key: https://console.aihubmix.com/?utm_source=llms-agent&utm_medium=model-llms

> **This model was retired on 2026-09-02.** Requests return 404 model_retired.
>
> The specs and pricing below stay readable for migration and cost comparison — this id is no longer a live integration target. Responses for this id also carry RFC 8594 headers: `Sunset` (the retirement instant), `Deprecation`, and `Link: …; rel="successor-version"` when a successor is published — read those to detect this automatically. All retirements: https://aihubmix.com/models/retirements

> V3 Ultra-Fast Version,The current price is a limited-time 50% discount and will return to the original price on July 31st. The original price is: input: $0.55/M, output: $2.2/M. The model provider is the Sophnet platform. DeepSeek V3 Fast is a high-TPS, ultra-fast version of DeepSeek V3 0324, featuring full-precision (non-quantized) performance, enhanced code and math capabilities, and faster responses!

DeepSeek V3 0324 is a powerful Mixture-of-Experts (MoE) model with a total parameter count of 671B, activating 37B parameters per token.
It adopts Multi-Head Latent Attention (MLA) and the DeepSeekMoE architecture to achieve efficient inference and economical training costs.
It innovatively implements a load balancing strategy without auxiliary loss and sets multi-token prediction training targets to enhance performance.
The model is pre-trained on 14.8 trillion diverse, high-quality tokens and further optimized through supervised fine-tuning and reinforcement learning stages to fully realize its capabilities.
Comprehensive evaluations show that DeepSeek V3 outperforms other open-source models and rivals leading closed-source models in performance.
The entire training process only requires 2.788M H800 GPU hours and remains highly stable, with no irrecoverable loss spikes or rollbacks.

> Capability flags (tool use, structured output, vision, reasoning, …) are **not published** for this model. AIHubMix lists them only after official confirmation and this id is not covered yet — their absence below means unverified, not unsupported. Everything else here comes from the live catalog. Verify with a minimal real call before relying on a capability (see https://aihubmix.com/agents.md — an HTTP 200 alone is not success).

- Developer: DeepSeek
- Context window: 32,000 tokens
- Input modalities: text
- Pricing: $0.56/M input tokens, $2.24/M output tokens

## Endpoints (base URL: https://aihubmix.com)

- `POST /v1/chat/completions` — OpenAI Chat Completions (`Authorization: Bearer $AIHUBMIX_API_KEY`)

## Example

```bash
curl -s https://aihubmix.com/v1/chat/completions \
  -H "Authorization: Bearer $AIHUBMIX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"DeepSeek-V3-Fast","messages":[{"role":"user","content":"Hello"}]}'
```

## Response

This id is retired, so the call above no longer succeeds. It returns HTTP 404:

```json
{"error":{"code":"model_retired","message":"The model DeepSeek-V3-Fast has been retired and is no longer available. (tid: …)","type":"Aihubmix_api_error"}}
```

The request shape is kept above so it can be diffed against a successor's — point `"model"` at a live id to get a real completion. Discover live ids in the catalog (no auth): https://aihubmix.com/api/v1/models?type=llm&sort_by=order

## Errors

Error responses carry a `tid` (trace id) — include it when contacting support. Reference: https://docs.aihubmix.com/en/FAQs/HTTP-Codes.md

- 400 — parameter error; most are passed through from the upstream provider (media: `prompt_missing`, `size_not_supported`, `n_not_within_range`, …)
- 401 — missing `Authorization` header, or the key is invalid/expired
- 403 — `insufficient_user_quota` (top up at https://console.aihubmix.com/?utm_source=llms-agent&utm_medium=model-llms), account suspended, or this key is not allowed to use this model
- 429 — rate limited; back off and retry
- 503 — no channel can serve the request (check the model id and your access), or the upstream provider is throttling; retry later

## More

- Model page: https://aihubmix.com/model/DeepSeek-V3-Fast
- Try in browser: https://playground.aihubmix.com/?model=DeepSeek-V3-Fast
- Compare with another model (human-facing, side-by-side specs and pricing): https://aihubmix.com/compare — pick this model and a peer there; published pairs are listed in https://aihubmix.com/sitemap-compare.xml, unpublished pairs 404 so do not compose the path by hand
- Full parameter schema: follow `https://aihubmix.com/model-data/index.json` — find this id and fetch its `path` (filenames are content-addressed; do not compose them by hand)
- Generate runnable code programmatically: npm `@aihubmix/codegen` — the generator behind the Playground's "Get Code" (4 protocols × 7 languages, media endpoints included); the body it builds is the exact wire body the Playground sends, so generated snippets and real requests cannot diverge. `@aihubmix/model-schema` (npm) translates the parameter schema above into codegen input
- Site index for agents: https://aihubmix.com/llms.txt · Onboarding: https://aihubmix.com/agents.md

---

Canonical version of this document: https://aihubmix.com/model/DeepSeek-V3-Fast/llms.txt