119 lines
4.6 KiB
TOML
119 lines
4.6 KiB
TOML
# Anthropic direct
|
|
[[model]]
|
|
id = "claude-opus-4-8"
|
|
provider = "anthropic"
|
|
context_window = 1000000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", vision = true, prompt_caching = { kind = "explicit", max_breakpoints = 4 } }
|
|
|
|
[[model]]
|
|
id = "claude-sonnet-4-6"
|
|
provider = "anthropic"
|
|
context_window = 1000000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "budget_tokens", vision = true, prompt_caching = { kind = "explicit", max_breakpoints = 4 } }
|
|
|
|
[[model]]
|
|
id = "claude-haiku-4-5"
|
|
provider = "anthropic"
|
|
context_window = 200000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "budget_tokens", vision = true, prompt_caching = { kind = "explicit", max_breakpoints = 4 } }
|
|
|
|
# Ollama local (capability is router-ish / ollama handles its own models)
|
|
[[model]]
|
|
id = "llama3.3"
|
|
provider = "ollama-local"
|
|
context_window = 128000
|
|
|
|
[[model]]
|
|
id = "qwen3-coder"
|
|
provider = "ollama-local"
|
|
context_window = 256000
|
|
|
|
# Codex OAuth (ChatGPT backend via Responses API)
|
|
# Codex CLI's model catalog exposes these coding models with a 272k effective
|
|
# context window on this route, even when public API docs advertise larger
|
|
# model-family windows.
|
|
# Codex currently caps the ChatGPT-backed route at 272k even though the public
|
|
# GPT-6 family and GPT-5.6 Sol advertise a 1.05M context window. Keep both values
|
|
# so resolution and compaction use the effective backend limit without losing
|
|
# each underlying model capability.
|
|
[[model]]
|
|
id = "gpt-6-astra"
|
|
provider = "codex-oauth"
|
|
context_window = 1050000
|
|
max_context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
# https://developers.openai.com/api/docs/models/gpt-6.1-sol
|
|
# Tool calling requires Responses; reasoning effort supports low through max,
|
|
# but not none or minimal. Preserve the Codex route's effective limit above.
|
|
[[model]]
|
|
id = "gpt-6.1-sol"
|
|
provider = "codex-oauth"
|
|
context_window = 1050000
|
|
max_context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-6-luna"
|
|
provider = "codex-oauth"
|
|
context_window = 1050000
|
|
max_context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.6-sol"
|
|
provider = "codex-oauth"
|
|
context_window = 1050000
|
|
max_context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.6-luna"
|
|
provider = "codex-oauth"
|
|
context_window = 1050000
|
|
max_context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.5"
|
|
provider = "codex-oauth"
|
|
context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.4"
|
|
provider = "codex-oauth"
|
|
context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.4-mini"
|
|
provider = "codex-oauth"
|
|
context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "gpt-5.3-codex-spark"
|
|
provider = "codex-oauth"
|
|
context_window = 272000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
# OpenRouter
|
|
[[model]]
|
|
id = "anthropic/claude-opus-4.8"
|
|
provider = "openrouter"
|
|
context_window = 1000000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "budget_tokens", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "anthropic/claude-sonnet-4.6"
|
|
provider = "openrouter"
|
|
context_window = 1000000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "budget_tokens", vision = true, prompt_caching = { kind = "auto" } }
|
|
|
|
[[model]]
|
|
id = "openai/gpt-5.5"
|
|
provider = "openrouter"
|
|
context_window = 1050000
|
|
capability = { tool_calling = "parallel", structured_output = "json_schema", reasoning = "effort", vision = true, prompt_caching = { kind = "auto" } }
|