// compare
Compare
Which model, which provider, which price. Specs come from the vendors' own documentation — where we cannot source a number, we leave the row out rather than estimate it.
Model vs model
current generation
Opus 5 vs Sonnet 5 for coding
measured on our own billing
Opus vs Sonnet for coding (measured)
current flagships
GPT-6 Astra vs Claude Fable 5.1
flagship vs default
GPT-6 Astra vs Opus 5 — is 2× worth it?
bulk Flash vs Anthropic default
V4.1 Flash vs Sonnet 5 — cheap bulk vs Anthropic default
previous Fable vs default
prev gen
Fable 5 vs Opus 5 — is the flagship worth 2×?
previous id vs current Flash
prev gen
V4 Flash vs V4.1 Flash — which id to call
everyday driver
one side unavailable
Sonnet 5 vs Gemini 3.1 Pro
previous generation
one side unavailable
Opus 4.7 vs GPT-5 (previous generation)
Provider vs provider
also
Looking for an alternative to a specific provider?
Short reads on Anthropic direct, AWS Bedrock, Google Vertex AI, OpenRouter and Vercel AI Gateway — what each does well, and where we're cheaper.