DeepSeek V4 Flash 0423
DeepSeek V4 Flash 0423 is served by 16 companies at prices ranging from $0.02 to $0.44 per million input tokens, a 18.3× difference for identical weights.
- cheapest host · OpenInference
- $0.02
- price spread across hosts
- 18×
- companies serving it
- 16
- intelligence index
- 24.4
Is it worth it?
No cheaper model matches it on quality.
It beats gpt-oss-120b at the same quality band, for more money.
Every host, compared
The same weights, priced and configured differently by 16 companies. Precision matters: 4-bit builds measurably degrade output and cost the same.
| Host | $/1M in | $/1M out | Cached | Context | Precision | Uptime |
|---|---|---|---|---|---|---|
| OpenInferencecheapest | $0.024 | $1.54 | $0.024 | 1.05M | fp4 | 99.8% |
| Relace | $0.030 | $1.28 | $0.030 | 1.05M | fp4 | 100% |
| StreamLake | $0.083 | $0.165 | $0.017 | 1.02M | fp8 | 99.7% |
| DeepInfra | $0.090 | $0.180 | $0.018 | 1.05M | fp8 | 97.6% |
| GMICloud | $0.091 | $0.182 | $0.018 | 1.05M | fp8 | 100% |
| Venice | $0.097 | $0.193 | $0.020 | 1.00M | unknown | 99.5% |
| DigitalOcean | $0.098 | $0.196 | $0.020 | 1.05M | unknown | 100% |
| SiliconFlow | $0.130 | $0.280 | $0.028 | 1.05M | fp8 | 98.1% |
| Baidu | $0.140 | $0.280 | $0.028 | 1.05M | fp8 | 100% |
| AtlasCloud | $0.140 | $0.280 | $0.028 | 1.05M | fp4 | 94.8% |
| Parasail | $0.140 | $0.280 | $0.070 | 1.05M | fp8 | 100% |
| Alibaba | $0.150 | $0.300 | $0.030 | 1.00M | fp8 | 100% |
| Mancer 2 | $0.190 | $0.500 | n/a | 1.05M | fp8 | 98% |
| Azure | $0.210 | $0.560 | $0.031 | 1.05M | unknown | 99.9% |
| Wafer | $0.400 | $0.170 | $0.013 | 1.05M | unknown | 96.2% |
| Cloudflare | $0.440 | $1.32 | $0.014 | 384K | unknown | 92.7% |
3 of 16 hosts serve this model at 4-bit precision (OpenInference, Relace, AtlasCloud), at the same price as full-precision hosts. See why this matters.
What it costs to run
- Marketing emails
- $0.152
- per 1,000 · short brief in, one email out
- Support replies
- $0.170
- per 1,000 · ticket plus help-centre context
- Lead enrichment
- $0.312
- per 1,000 · company page in, fields out
- Long-form drafts
- $0.612
- per 1,000 · outline in, full draft out
- Questions on docs
- $1.13
- per 1,000 · one large document, many questions
- Agent / tool loop
- $0.910
- per 1,000 · long context reused each step
Priced on the median host. Change the workload on the leaderboard to see how the ranking moves.
Specification
- Context, most hosts
- 1.05M · 1,573 pages
- Context range across hosts
- 384K to 1.05M
- Max output
- 944K
- Cache discount
- 78% off cached input
- Input types
- text
- Tool use
- supported
- Reasoning
- optional
- Arena Elo
- 1172 across 8 categories
- Agentic Elo
- n/a
- AA Intelligence
- 24.4
Other models with the same host spread
Host pricing and precision read live from OpenRouter’s endpoint API, compiled 12 October 2026. Free to quote with attribution. See all findings.