guide

AWS maps three layers of Bedrock token-cost visibility

Rachanee Singprasong and Chitresh Saxena lay out three layers of Bedrock cost visibility: native CloudWatch metrics plus CUR 2.0 caller-identity attribution, then model invocation logging, then OpenTelemetry in the client.

Published 2026-06-29Source: Amazon Web Services

Why it matters

A billing export shows the total, not the developer, model or retry that produced it. CUR 2.0 can now attribute Bedrock spend to the IAM principal behind each individual call, and Claude Code ships OpenTelemetry support to cover the client side.

Tokenmaxxing read

AWS puts the savings at 30-50%, and it comes from ordinary routing discipline: Sonnet 4.5 at $3/$15 per million tokens against Haiku 4.5 at $1/$5 is roughly 67% off routine work, with cached input up to 90% cheaper. Telemetry just shows which tasks sit on the wrong model.

Source takeaway

Instrument the client, not Bedrock — that is where per-session cost and cache-hit ratios live. Coverage is uneven: Claude Code is the worked example, while Cursor shops are held to the first two layers plus that vendor's own admin dashboard.

Topic links

ai-spendtopicllm-observabilitycost-governancetopic model-routingtopic coding-agentstopic

Related projects

Tools that match this angle

#1Direct

Routing

LiteLLM

BerriAI/litellm

An OpenAI-compatible gateway and SDK for calling many model providers with budgets, logging, load balancing, guardrails, and cost tracking.

59.1K11.6KSource-available

gatewaycost-trackingrouting

Project profile GitHub

#2Direct

Observability

Langfuse

langfuse/langfuse

Open-source LLM engineering platform for observability, traces, metrics, evals, prompt management, datasets, and playground workflows.

34.8K3.8KSource-available

tracesevalscosts

Project profile GitHub

#4In spirit

Agents

LangGraph

langchain-ai/langgraph

A framework for building resilient stateful agents with explicit graphs, persistence, human-in-the-loop flows, and controllable execution.

41.9K7.1KMIT

agentsstateworkflows

Project profile GitHub

Related feed

More source-linked context

long-formEC

long-form2026-09-18

Enterprise AI budgets break at the handoff to production

Express Computer interviews New Relic India's Ganesh Narasimhadevara on why AI bills keep climbing while the blended cost per million tokens has fallen over a year, from $18.40 in Q1 2025 down to $6.07 in Q1 2026.

tokenmaxxingai-spendcost-governance

Read note

newsCN

news2026-09-15

Cisco's Splunk adds Tokenomics to track coding-agent token spend

At Splunk .conf on Sept. 15, Cisco added a Tokenomics module to Splunk Agent Observability. It attributes token spend across AI agents and across employees' use of coding agents, naming Claude Code, Codex and Cursor.

ai-spendcoding-agentsllm-observability

Read note

newsNT

news2026-08-11

Route AI agents across models with NVIDIA NeMo Switchyard

NVIDIA shipped NeMo Switchyard, a provider-agnostic SDK that escalates agent steps from cheap models to frontier ones only when a task demands it. LangChain benchmarked it over 145 multi-turn agentic tasks.

model-routingagentsai-spend

Read note