aws-samples / aws-samples/bedrock-chat

[FEATURE] Add Claude Opus 4.8 (anthropic.claude-opus-4-8) to the model list

Open
#1,124 0 comments 1 reaction 0 assignees View on GitHub
planned
Dominant language
TypeScript
Stars
1.3k
Forks
535
Avg merge
1d 12h
Merged PRs (30d)
10

Description

### What

Claude Opus 4.8 is live in Amazon Bedrock (`anthropic.claude-opus-4-8`, lifecycle ACTIVE), but bedrock-chat's model map currently tops out at `claude-v4.7-opus`. Adding it follows the same pattern as the Opus 4.7 addition (#1101): same request surface (adaptive thinking only, no assistant prefill, no temperature/top_p/top_k) and the same list pricing as 4.7 ($5/$25 per MTok, no long-context premium).

### Verified (live, from an eu-central-2 deployment)

- `aws bedrock list-foundation-models --region eu-central-2` → `anthropic.claude-opus-4-8` ACTIVE
- `aws bedrock list-inference-profiles --region eu-central-2` → `eu.anthropic.claude-opus-4-8` and `global.anthropic.claude-opus-4-8` ACTIVE
- `get-inference-profile` on the `eu.` profile → covers eu-central-1, eu-central-2, eu-north-1, eu-south-1, eu-south-2, eu-west-1, eu-west-3. **Note: eu-west-2 is NOT covered** (it is for 4.7), so copying the 4.7 region list would be wrong.

### Caveat

Our AWS org restricts us to EU regions, so the us/jp/au/apac entries of the inference maps can't be verified from here. The PR I'm submitting keeps those geos out and ships only the verified EU + global entries; maintainers can extend the maps by running `scripts/cross_region_inference/get_supported_cross_region_inferences.py` from an unrestricted account.

PR follows, mirroring #1101 file-by-file.

Contributor guide

Open the contributing guide

Research direction

Compare the model map and inference maps with the existing Claude Opus 4.7 addition in issue #1101. Use scripts/cross_region_inference/get_supported_cross_region_inferences.py to verify supported regions, then confirm the new model exposes the verified EU and global entries without adding eu-west-2; done means the model list and inference mappings include anthropic.claude-opus-4-8.

Written by the indexing model from the issue text.

Assessment

Tech stack
aws, typescript
Domain
backend, cloud
Issue type
Feature
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
72/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.