model manifest

> Devin (agent)

Name: Devin (agent)
Brand: Cognition

by Cognition

in: textout: textOther

prompt /1M

—

completion /1M

—

providers

share this modelX LinkedIn Reddit

// specification— means undocumented, never guessed

family: Other
modalities: text → text(input → output · legacy primary: text)
tokenizer: —
max output tokens: —
knowledge cutoff: —
license: —
spec / docs: https://devin.ai
added to index: 1d ago

// aboutsourced from the operator docs

Autonomous software-engineering agent — priced per ACU, not per token

source: https://devin.ai

// capabilitiescheckmarks only when documented — unknown stays unknown

Streaming

Tool calling

Structured outputs

Image input

Audio input

Reasoning

Image output

// providers_&_endpoints1 endpoint · sorted by input price

Provider	In /1M	Out /1M	Context	Quant	Connect	Uptime 7d	p50 / p95	Tok/s
DevinSubscription router · verified 1d agoprice source	—	—	—	—	oauth+ connect	—	—	—

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

No sourced benchmark results recorded for this model yet — scores appear here only with a source, evaluation conditions and dates. Nothing is estimated.

// call_this_model

curl https://www.uprouter.online/api/connect/v1/chat/completions \
  -H "Authorization: Bearer upr_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "devin-agent",
    "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 devin-agent slug.

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