custom @ai-sdk/anthropic provider , context token usage display error

[caozhiyuan]

Description

{"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null},"usage":{"input_tokens":236,"output_tokens":81,"cache_read_input_tokens":12800}}

	"provider": {
		"local": {
			"npm": "@ai-sdk/anthropic",
			"name": "My Local",
			"options": {
				"baseURL": "http://localhost:4141/v1",
				"apiKey": "dummy"
			},
			"models": {
				"gpt-5.2": {
					"name": "gpt-5.2",
					"limit": {
						"context": 200000,
						"output": 64000
					}
				},
				"gpt-5-mini": {
					"name": "gpt-5-mini",
					"limit": {
						"context": 200000,
						"output": 64000
					}
				},
				"claude-sonnet-4.5": {
					"id": "claude-sonnet-4.5",
					"name": "claude-sonnet-4.5",
					"limit": {
						"context": 200000,
						"output": 16000
					},
					"options": {
						"thinking": {
							"type": "enabled",
							"budgetTokens": 15999
						}
					}
				}
			}
		}
	}

Plugins

No response

OpenCode version

1.1.8

Steps to reproduce

1. ask hi
2. see context info

Screenshot and/or share link

No response

Operating System

windows 11

Terminal

pwsh

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• Over 100% context usage #7025: Over 100% context usage - relates to context token display calculation errors
• Custom OpenAI-compatible provider options not being passed to API calls #5674: Custom OpenAI-compatible provider options not being passed to API calls - directly related to custom provider configuration issues with options like baseURL and apiKey
• Partial thinking config doesn't merge with defaults, causing AI_InvalidArgumentError #6806: Partial thinking config doesn't merge with defaults - related to custom @ai-sdk/anthropic provider configuration with incomplete options
• Anthropic team plan statistics don't show usage while in opencode #7231: Anthropic team plan statistics don't show usage in OpenCode - relates to Anthropic provider usage and token reporting issues

Feel free to ignore if none of these address your specific case.

[caozhiyuan]

seems https://github.com/vercel/ai/blob/ai%405.0.97/packages/anthropic/src/anthropic-messages-language-model.ts

[caozhiyuan]

@rekram1-node PR in #7643 , please help review.

[caozhiyuan]

@rekram1-node https://github.com/vercel/ai/blob/ai%405.0.97/packages/anthropic/src/anthropic-messages-language-model.ts#L1ai269
ai sdk anthropic set inputTokens and cachedInputTokens at message_start.
If inputTokens is 0, should taking from rawUsage at message_stop https://github.com/vercel/ai/blob/ai%405.0.97/packages/anthropic/src/anthropic-messages-language-model.ts#L1329.