DeepSeek V4 Flash 0423
✓ VERIFIED BY DEEPSEEKDeepSeek · Open weights · Released Apr 24, 2026 · deepseek/deepseek-v4-flash
Best $0.014 in / $0.029 out via Relace · fp4
1.05M contextToolsReasoningStructured outputPrompt cachingOpen weights
Available from 15 providersUSD / MTOK · PRICES 15 MINUTES AGO · SPEED 10 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | SMOKE | UPTIME |
|---|---|---|---|---|---|---|---|---|---|
| $0.014 | $0.45 | $0.014 | 1.05M | fp4 | 0.65s° | 37° tok/s | — | 100.0% | |
| $0.015 | $0.029 | $0.003 | 1.05M | fp8 | 1.1s° | 131° tok/s | 5/5 | 99.7% | |
| $0.028 | $0.056 | $0.006 | 1.02M | fp8 | 3.9s° | 43° tok/s | 5/5 | 99.0% | |
| $0.07 | $0.14 | $0.014 | 1.05M | — | 0.64s° | 181° tok/s | — | 98.3% | |
| $0.072 | $0.144 | $0.014 | 1.05M | fp8 | — | — | 5/5 | 100.0% | |
| $0.09 | $0.18 | $0.018 | 1.05M | fp8 | 0.8s° | 34° tok/s | — | 100.0% | |
| $0.12 | $0.17 | $0.024 | 1.05M | mxfp4 | 0.37s° | 211° tok/s | — | 100.0% | |
| $0.13 | $0.28 | $0.028 | 1.05M | fp8 | 0.71s° | 93° tok/s | — | 99.4% | |
| $0.135 | $0.27 | $0.027 | 1.05M | fp8 | — | — | — | 100.0% | |
| $0.138 | $0.275 | $0.028 | 1M | — | 1.2s° | 71° tok/s | — | 99.9% | |
| $0.14 | $0.28 | $0.07 | 1.05M | fp8 | 0.81s° | 113° tok/s | — | 99.9% | |
| $0.18 | $0.50 | $0.07 | 1.05M | fp8 | 0.73s° | 95° tok/s | — | 99.9% | |
| $0.21 | $0.56 | $0.031 | 1.05M | — | ✕ | ✕ | — | 99.7% | |
| $0.27 | $1.08 | $0.027 | — | — | ✕ | ✕ | — | — | |
| $0.44 | $1.32 | $0.014 | 1M | — | 0.94s | 138 tok/s | — | — |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured against the provider API directly; ° = routed via OpenRouter (adds a hop) · ✕ = probed, endpoint returned no measurable stream · SMOKE: of 5 programmatic capability probes passed (JSON schema, instruction following, tool call, long-context, benign compliance; retry-once) · screened via OpenRouter routing · v0.2 · methodology v1.1 · measured 10 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Quality benchmarksCURATED AUG 24, 2026
Third-party reported results, hand-curated with sources — not measured by ModelIndex. Our own measured serving data (TTFT/TPS/SMOKE) is in the providers table above.
Input price history · best listed price, observed
↓ 77% SINCE SEP 22, 2026
$0.06 · SEP 22, 2026$0.014 · TODAY
Pricing variants · USD / MTok in / out
STANDARD
$0.014 / $0.45
per MTok in / out
CACHE READ $0.014CACHE WRITE Free
Recent changesRSS ↗
Nov 21, 2026endpoint removed · atlascloud17 → 15
Nov 21, 2026endpoint removed · gmicloud17 → 15
Nov 20, 2026price decreased · output0.1652 → 0.14
Nov 16, 2026price increased · input0.014 → 0.05
Nov 16, 2026price increased · output0.0294 → 0.1652
Nov 14, 2026price decreased · input0.05 → 0.0307
Nov 14, 2026price decreased · output0.1652 → 0.0624
Nov 10, 2026endpoint removed · digitalocean16 → 15
About
DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...
Specifications
Context window1,048,576
Max output393,216
Modalities intext
Modalities outtext
LicenseOpen weights
ReleasedApr 24, 2026
TokenizerDeepSeek
WeightsHugging Face ↗
Related
Embed badge
[](https://modelindex.ai/models/deepseek/deepseek-v4-flash)Sources
deepseek ↗15 minutes ago
OFFICIAL PRICINGOpenRouter API15 minutes ago
AGGREGATORLiteLLM dataset15 minutes ago
AGGREGATORHugging Face15 minutes ago
COMMUNITYModelIndex benchmarks18 days ago
MEASURED✓ Verified against DeepSeek’s own source.
Verified 15 minutes ago · list prices, not negotiated rates.