UPROUTERONLINE

model manifest

> deepseek-v4-pro

by deepseek

in: textout: textctx: 1.0MDeepSeek
prompt /1M
$0.698
completion /1M
$1.40
providers
12
Tool callingStructured outputsReasoning
// specification— means undocumented, never guessed
family
DeepSeek
modalities
texttext(input → output · legacy primary: text)
tokenizer
DeepSeek
max output tokens
knowledge cutoff
license
added to index
9d ago
// aboutsourced from the operator docs

Offered free via imported source data (freeairouter.com).

source: https://openrouter.ai/deepseek/deepseek-v4-pro

// capabilitiescheckmarks only when documented — unknown stays unknown
Streaming undocumented
Tool calling supported
Structured outputs supported
Image input undocumented
Audio input undocumented
Reasoning supported
Image output undocumented

supported sampling parameters

frequency_penaltylogprobsmax_completion_tokensmax_tokenspresence_penaltyrepetition_penaltyresponse_formatseedstoptemperaturetop_ktop_logprobstop_p
// providers_&_endpoints12 endpoints · sorted by input price
ProviderIn /1MOut /1MContextQuantConnectUptime 7dp50 / p95Tok/s
Charm HyperPaid API · verified 4d agoprice source $0.0000$0.00001.0MPooled
FreebuffFree relay · verified 4d agoprice source $0.0000$0.00001.0M
Naga.acCommercial aggregator · verified 4d agoprice source $0.0000$0.00001.0MPooled
OpenAdapterPaid API · verified 4d agoprice source $0.0000$0.00001.0MPooled
OpenVectaOfficial free tier · verified 4d agoprice source $0.0000$0.00001.0MPooled
RoutewayPaid API · verified 4d agoprice source $0.0000$0.00001.0MBYOK
Zylo APIPaid API · verified 4d agoprice source $0.0000$0.00001.0MPooled
Lightning AIOfficial free tier · verified 12d agoprice source 1.0MPooled100%89 / 1862 ms
SupXH FreeFree relay · verified 12d agoprice source 1.0MPooled100%299 / 602 ms
TGhanCommercial aggregator · verified 12d agoprice source 1.0MPooled100%685 / 1360 ms
NVIDIA NIMOfficial free tierprice source BYOK100%15 / 370 ms
VSLLMFree relayprice source Pooled100%224 / 392 ms

Per-endpoint prices are normalized from each provider’s public pricing page (verify before purchasing). Uptime/latency derive from our own probes of the provider’s API host — they are provider-level, not model-specific. Throughput is shown only when actually measured; absent means not measured, never estimated.

// benchmarkscomparable only within matching conditions

Artificial Analysis Agentic Index

communitycommunity-reported
27.9 score
verified 4d agosource

conditions: artificialanalysis.ai agentic composite (aggregated via OpenRouter model listing)

Artificial Analysis Coding Index

communitycommunity-reported
59.4 score
verified 4d agosource

conditions: artificialanalysis.ai coding composite (aggregated via OpenRouter model listing)

SWE-bench Verified

0813independent evalcommunity-reported
96.4%

conditions: Vals AI evaluation harness

Scores are comparable only within matching evaluation conditions (scaffold, agent vs single-shot, split/version). Different benchmarks and versions are not directly comparable to each other.

// call_this_model
curl https://uprouter.online/api/connect/v1/chat/completions \
  -H "Authorization: Bearer upr_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-pro",
    "messages": [{ "role": "user", "content": "Hello via Uprouter Connect" }]
  }'

Metered in Uprouter compute — stage a provider above and add your key on the Connect dashboard; the gateway routes the deepseek-v4-pro slug.

This manifest records only what the operator documents — unknown fields stay unknown. ← back to the model catalog

Uprouter command palette

Jump to a page, provider, model or action

Keyboard shortcuts

Every keyboard shortcut available on Uprouter