DeepSeek V4 Pro 0423
DeepSeek V4 Pro 0423 is served by 15 companies at prices ranging from $0.45 to $1.91 per million input tokens, a 4.2× difference for identical weights.
- cheapest host · Parasail
- $0.45
- price spread across hosts
- 4.2×
- companies serving it
- 15
- intelligence index
- 30.4
Is it worth it?
Beaten on both axes. GLM 5.3 Flash scores 41.8 against 30.4 and costs 7.3× less for the same work.
Every host, compared
The same weights, priced and configured differently by 15 companies. Precision matters: 4-bit builds measurably degrade output and cost the same.
| Host | $/1M in | $/1M out | Cached | Context | Precision | Uptime |
|---|---|---|---|---|---|---|
| Parasailcheapest | $0.450 | $3.48 | $0.100 | 1.05M | fp8 | 97.5% |
| StreamLake | $0.955 | $1.91 | $0.080 | 1.02M | fp8 | 100% |
| GMICloud | $0.957 | $1.91 | $0.080 | 1.05M | fp8 | 99.6% |
| DigitalOcean | $1.04 | $2.09 | $0.209 | 1.05M | unknown | 100% |
| Reka | $1.05 | $10.50 | $0.210 | 1.05M | unknown | 98.7% |
| Cloudflare | $1.15 | $2.55 | $0.200 | 1.05M | unknown | 99.6% |
| DeepInfra | $1.30 | $2.60 | $0.100 | 1.05M | fp8 | 100% |
| Alibaba | $1.42 | $2.83 | $0.118 | 1.00M | fp8 | 96.3% |
| SiliconFlow | $1.50 | $3.14 | $0.135 | 1.05M | fp8 | 99.9% |
| Alibaba | $1.56 | $3.12 | $0.130 | 1.00M | fp8 | 98.9% |
| Novita | $1.60 | $3.20 | $0.135 | 1.05M | fp8 | 100% |
| Venice | $1.65 | $3.30 | $0.330 | 1.00M | unknown | 99.7% |
| AtlasCloud | $1.68 | $3.38 | $0.130 | 1.05M | fp4 | 99.9% |
| Baidu | $1.69 | $3.38 | $0.140 | 1.05M | fp8 | 100% |
| Azure | $1.91 | $3.83 | $0.160 | 1.05M | unknown | 100% |
1 of 15 hosts serve this model at 4-bit precision (AtlasCloud), at the same price as full-precision hosts. See why this matters.
What it costs to run
- Marketing emails
- $1.66
- per 1,000 · short brief in, one email out
- Support replies
- $1.68
- per 1,000 · ticket plus help-centre context
- Lead enrichment
- $3.30
- per 1,000 · company page in, fields out
- Long-form drafts
- $6.78
- per 1,000 · outline in, full draft out
- Questions on docs
- $8.19
- per 1,000 · one large document, many questions
- Agent / tool loop
- $6.71
- 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
- 1.00M to 1.05M
- Max output
- 384K
- Cache discount
- 94% off cached input
- Input types
- text
- Tool use
- supported
- Reasoning
- optional
- Arena Elo
- 1212 across 8 categories
- Agentic Elo
- 1002
- AA Intelligence
- 30.4
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.