Requesty
Fireworks AI

deepseek-v4-flash

DeepSeek V4 Flash is a streamlined open-source Mixture-of-Experts model (284B parameters) optimized for fast, cost-efficient inference while preserving strong reasoning and coding performance at 1M token context scale. It leverages hybrid attention innovations for lower latency and higher throughput, delivering near-Pro reasoning quality ideal for interactive agents and high-volume production workloads.

ReasoningTool callingCaching
CompareDocs

Specifications

Context window1M tokens
Max output131K tokens
API typechat
AddedMay 27, 2026
Model IDfireworks/deepseek-v4-flash
Data retentionNo
Used for trainingNo
Provider location🇺🇸 US
Privacy policyFireworks Privacy Policy→

Benchmarks

Released 2026-04-24
Coding Indexcoding
56.2%

Artificial Analysis Coding Index — a composite of coding evaluations including LiveCodeBench, SciCode and Terminal-Bench.

GPQA Diamondreasoning
89.4%

Graduate-level physics, chemistry & biology questions designed to resist Googling.

Intelligence Indexreasoning
40.3%

Artificial Analysis Intelligence Index — a composite of multiple evaluations measuring overall model capability.

Scores are sourced from official model cards, Artificial Analysis, and public leaderboards. Benchmarks measure specific skills and do not capture every aspect of model quality. Always test on your own workload.

Pricing

Input / 1M
$0.14
Output / 1M
$0.28
Cache write
N/A
Cache read / 1M
$0.03
Estimated cost
100K input + 10K output$0.0168
1M input + 100K output$0.17
10M input + 1M output$1.68

Requesty charges exactly what the upstream provider charges, no markup, no per-request fees. Prompt caching and smart routing can reduce effective cost by 30-80%.

Quickstart

Drop-in compatible with the OpenAI SDK. Change the base URL, swap in your Requesty API key, and set the model to fireworks/deepseek-v4-flash.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
)

response = client.chat.completions.create(
    model="fireworks/deepseek-v4-flash",
    messages=[
        {"role": "user", "content": "Explain quantum computing in one paragraph."},
    ],
)

print(response.choices[0].message.content)

Other Fireworks AI models

glm-5.2
1M ctx$1.40/1M
Coding 69
qwen3.7-plus
262K ctx$0.40/1M
Coding 56
nemotron-3-ultra-nvfp4
262K ctx$0.60/1M
minimax-m3
512K ctx$0.30/1M
Coding 59
kimi-k2.7-code
262K ctx$0.95/1M
Coding 61
gpt-oss-120b
131K ctx$0.15/1M
Coding 30

Frequently asked questions

How much does deepseek-v4-flash cost?
deepseek-v4-flash is priced at $0.14 per million input tokens and $0.28 per million output tokens when accessed via Requesty. Prompt caching is supported, which can cut effective input cost by up to 90% on repeated context. Requesty charges exactly what the upstream provider charges, we don't add markup.
What is the context window of deepseek-v4-flash?
deepseek-v4-flash has a context window of 1M tokens, with a maximum output of 131K tokens per response. That's roughly 1,333 words of input you can fit in a single prompt.
How does deepseek-v4-flash perform on benchmarks?
deepseek-v4-flash scores 95.0% on τ²-Bench, 89.4% on GPQA Diamond, 56.2% on Coding Index. See the full benchmark chart above for results across MMLU Pro, GPQA Diamond, SWE-Bench Verified, HumanEval, MATH, AIME, MMMU, and LiveBench.
What can deepseek-v4-flash do?
deepseek-v4-flash supports tool calling, extended reasoning, prompt caching. You can call it through any OpenAI-compatible client by pointing base_url to Requesty.
How do I use deepseek-v4-flash with the OpenAI SDK?
Install the OpenAI SDK, set base_url to "https://router.requesty.ai/v1", set your API key to your Requesty key, and set the model to "fireworks/deepseek-v4-flash". The Quickstart above shows Python, JavaScript and cURL snippets.
Can I run deepseek-v4-flash through Requesty?
Yes. deepseek-v4-flash runs through Requesty's OpenAI-compatible API, served from Fireworks AI. You do not host the model yourself: point base_url at Requesty, set the model to "fireworks/deepseek-v4-flash", and requests are routed to the upstream provider with automatic failover. The same key gives you 400+ other models too.

Access deepseek-v4-flash through Requesty

One API key, 400+ models, OpenAI-compatible. No markup on provider prices, automatic failover, and smart caching built-in.

All Fireworks AI models