Summary
Claude Sonnet 4.6 (claude-sonnet-4-6) has max_output_tokens: 64000 in the catalog, but Anthropic's official documentation states the max output is 128,000 tokens. This affects the primary Anthropic API entry and all provider-specific entries (Bedrock and Vertex).
The 128k output limit is confirmed on both the models overview page (comparison table: "Max output: 128k tokens") and the pricing page (long-context section lists Sonnet 4.6 alongside other 128k-output models). This is consistent with Claude Opus 4.8, Opus 4.7, and Opus 4.6 — all of which already have max_output_tokens: 128000 in the catalog.
Affected Entries
| Model ID |
Provider |
Current max_output_tokens |
Correct Value |
Line |
claude-sonnet-4-6 |
anthropic |
64000 |
128000 |
~1638 |
anthropic.claude-sonnet-4-6 |
bedrock |
64000 |
128000 |
~6804 |
publishers/anthropic/models/claude-sonnet-4-6 |
vertex |
64000 |
128000 |
~8185 |
Two additional Bedrock regional entries also need the same update:
us.anthropic.claude-sonnet-4-6 (line ~6768) — 64000 → 128000
global.anthropic.claude-sonnet-4-6 (line ~6786) — 64000 → 128000
All other fields (pricing, format, reasoning, etc.) are correct and unchanged.
Verification
| Check |
Status |
Detail |
| Cross-source (max_output_tokens) |
✅ |
Models overview: comparison table says "Max output: 128k tokens" AND Pricing page: long-context section lists Sonnet 4.6 alongside other 128k-output models |
| Already fixed |
✅ |
Confirmed all 5 entries still have max_output_tokens: 64000 in model_list.json |
| Sibling consistency |
✅ |
Claude Opus 4.8, Opus 4.7, and Opus 4.6 all already have max_output_tokens: 128000 in the catalog |
| Duplicate check |
✅ |
No open issue covers Claude Sonnet 4.6 max_output_tokens |
Verification Notes
| Field |
Source |
Notes |
| Max output = 128k tokens |
Models overview |
Comparison table explicitly says "128k tokens" for Claude Sonnet 4.6 |
| Max output = 128k tokens |
Pricing page |
"Claude Fable 5, Claude Opus 4.8, Opus 4.7, Opus 4.6, and Sonnet 4.6 include the full 1M token context window at standard pricing" — grouped with other 128k-output models |
| 300k via Batch API |
Models overview |
"On the Message Batches API, Claude Opus 4.8, Opus 4.7, Opus 4.6, and Sonnet 4.6 support up to 300k output tokens" — not relevant to max_output_tokens field |
Local Files Inspected
packages/proxy/schema/model_list.json — all 5 Sonnet 4.6 entries confirmed at 64000
- Claude Sonnet 4.5 (
claude-sonnet-4-5) correctly has max_output_tokens: 64000 — this is NOT affected (Sonnet 4.5 docs confirm 64k)
{
"kind": "stale_metadata",
"provider": "anthropic",
"models": ["claude-sonnet-4-6", "anthropic.claude-sonnet-4-6", "publishers/anthropic/models/claude-sonnet-4-6"],
"status": "active",
"model_specs": {
"claude-sonnet-4-6": {
"format": "anthropic",
"flavor": "chat",
"multimodal": true,
"input_cost_per_mil_tokens": 3,
"output_cost_per_mil_tokens": 15,
"input_cache_read_cost_per_mil_tokens": 0.3,
"input_cache_write_cost_per_mil_tokens": 3.75,
"input_cache_write_5m_cost_per_mil_tokens": 3.75,
"input_cache_write_1h_cost_per_mil_tokens": 6,
"displayName": "Claude Sonnet 4.6",
"reasoning": true,
"reasoning_budget": true,
"fallback_models": [
"anthropic.claude-sonnet-4-6",
"global.anthropic.claude-sonnet-4-6",
"publishers/anthropic/models/claude-sonnet-4-6",
"us.anthropic.claude-sonnet-4-6"
],
"max_input_tokens": 1000000,
"max_output_tokens": 128000,
"available_providers": ["anthropic"]
},
"anthropic.claude-sonnet-4-6": {
"format": "anthropic",
"flavor": "chat",
"multimodal": true,
"input_cost_per_mil_tokens": 3,
"output_cost_per_mil_tokens": 15,
"input_cache_read_cost_per_mil_tokens": 0.3,
"input_cache_write_cost_per_mil_tokens": 3.75,
"displayName": "Claude Sonnet 4.6",
"reasoning": true,
"reasoning_budget": true,
"max_input_tokens": 1000000,
"max_output_tokens": 128000,
"available_providers": ["bedrock"]
},
"publishers/anthropic/models/claude-sonnet-4-6": {
"format": "anthropic",
"flavor": "chat",
"multimodal": true,
"input_cost_per_mil_tokens": 3,
"output_cost_per_mil_tokens": 15,
"input_cache_read_cost_per_mil_tokens": 0.3,
"input_cache_write_cost_per_mil_tokens": 3.75,
"input_cache_write_5m_cost_per_mil_tokens": 3.75,
"input_cache_write_1h_cost_per_mil_tokens": 6,
"displayName": "Claude Sonnet 4.6",
"reasoning": true,
"reasoning_budget": true,
"max_input_tokens": 1000000,
"max_output_tokens": 128000,
"available_providers": ["vertex"]
}
},
"source_urls": [
"https://platform.claude.com/docs/en/docs/about-claude/models/all-models",
"https://platform.claude.com/docs/en/about-claude/pricing"
]
}
Summary
Claude Sonnet 4.6 (
claude-sonnet-4-6) hasmax_output_tokens: 64000in the catalog, but Anthropic's official documentation states the max output is 128,000 tokens. This affects the primary Anthropic API entry and all provider-specific entries (Bedrock and Vertex).The 128k output limit is confirmed on both the models overview page (comparison table: "Max output: 128k tokens") and the pricing page (long-context section lists Sonnet 4.6 alongside other 128k-output models). This is consistent with Claude Opus 4.8, Opus 4.7, and Opus 4.6 — all of which already have
max_output_tokens: 128000in the catalog.Affected Entries
max_output_tokensclaude-sonnet-4-6anthropic.claude-sonnet-4-6publishers/anthropic/models/claude-sonnet-4-6Two additional Bedrock regional entries also need the same update:
us.anthropic.claude-sonnet-4-6(line ~6768) — 64000 → 128000global.anthropic.claude-sonnet-4-6(line ~6786) — 64000 → 128000All other fields (pricing, format, reasoning, etc.) are correct and unchanged.
Verification
max_output_tokens: 64000inmodel_list.jsonmax_output_tokens: 128000in the catalogVerification Notes
max_output_tokensfieldLocal Files Inspected
packages/proxy/schema/model_list.json— all 5 Sonnet 4.6 entries confirmed at 64000claude-sonnet-4-5) correctly hasmax_output_tokens: 64000— this is NOT affected (Sonnet 4.5 docs confirm 64k){ "kind": "stale_metadata", "provider": "anthropic", "models": ["claude-sonnet-4-6", "anthropic.claude-sonnet-4-6", "publishers/anthropic/models/claude-sonnet-4-6"], "status": "active", "model_specs": { "claude-sonnet-4-6": { "format": "anthropic", "flavor": "chat", "multimodal": true, "input_cost_per_mil_tokens": 3, "output_cost_per_mil_tokens": 15, "input_cache_read_cost_per_mil_tokens": 0.3, "input_cache_write_cost_per_mil_tokens": 3.75, "input_cache_write_5m_cost_per_mil_tokens": 3.75, "input_cache_write_1h_cost_per_mil_tokens": 6, "displayName": "Claude Sonnet 4.6", "reasoning": true, "reasoning_budget": true, "fallback_models": [ "anthropic.claude-sonnet-4-6", "global.anthropic.claude-sonnet-4-6", "publishers/anthropic/models/claude-sonnet-4-6", "us.anthropic.claude-sonnet-4-6" ], "max_input_tokens": 1000000, "max_output_tokens": 128000, "available_providers": ["anthropic"] }, "anthropic.claude-sonnet-4-6": { "format": "anthropic", "flavor": "chat", "multimodal": true, "input_cost_per_mil_tokens": 3, "output_cost_per_mil_tokens": 15, "input_cache_read_cost_per_mil_tokens": 0.3, "input_cache_write_cost_per_mil_tokens": 3.75, "displayName": "Claude Sonnet 4.6", "reasoning": true, "reasoning_budget": true, "max_input_tokens": 1000000, "max_output_tokens": 128000, "available_providers": ["bedrock"] }, "publishers/anthropic/models/claude-sonnet-4-6": { "format": "anthropic", "flavor": "chat", "multimodal": true, "input_cost_per_mil_tokens": 3, "output_cost_per_mil_tokens": 15, "input_cache_read_cost_per_mil_tokens": 0.3, "input_cache_write_cost_per_mil_tokens": 3.75, "input_cache_write_5m_cost_per_mil_tokens": 3.75, "input_cache_write_1h_cost_per_mil_tokens": 6, "displayName": "Claude Sonnet 4.6", "reasoning": true, "reasoning_budget": true, "max_input_tokens": 1000000, "max_output_tokens": 128000, "available_providers": ["vertex"] } }, "source_urls": [ "https://platform.claude.com/docs/en/docs/about-claude/models/all-models", "https://platform.claude.com/docs/en/about-claude/pricing" ] }