CLAUDE LABJP
MODEL — Claude Opus 4.8 lands, improving coding, agentic, and reasoning over 4.7 at the same priceCODE — Opus 4.8's Fast mode runs at 2.5x speed and is now three times cheaper than earlier modelsCODE — Auto-mode command classification expands, with denial tracking and live bash path autocompleteENTERPRISE — Connector permissions in custom roles let admins control which tools each role can useTEAM — Tag Claude directly in Slack and hand off tasks while you focus elsewhereMCP — MCP servers now show startup auth notices, making connection status easier to trackMODEL — Claude Opus 4.8 lands, improving coding, agentic, and reasoning over 4.7 at the same priceCODE — Opus 4.8's Fast mode runs at 2.5x speed and is now three times cheaper than earlier modelsCODE — Auto-mode command classification expands, with denial tracking and live bash path autocompleteENTERPRISE — Connector permissions in custom roles let admins control which tools each role can useTEAM — Tag Claude directly in Slack and hand off tasks while you focus elsewhereMCP — MCP servers now show startup auth notices, making connection status easier to track
TAG

tool_result

1 articles
Back to all tags
Related:
Claude API1tool use1vision1token optimization1
API & SDK/2026-06-29Advanced

Let Claude Actually See the Images Your Tools Return — Use Image Blocks in tool_result and Cut Tokens by Roughly 10x

Stuffing a base64 string into a tool_result makes the same image cost roughly 10–20x more tokens. Here is how to return it as an image content block instead, with SDK code, a token-cost estimate, and the gotchas I hit in production.