Model catalogue

Anthropic's most capable model. Top of SWE-bench Verified and top of the creative-writing arena — the one to reach for when being right matters more than what it costs.

Complex agentic coding and long-horizon work at half the price of Fable 5. Leads Terminal-Bench 2.1 and the maths arena.

Near-Opus coding quality at Sonnet money — 0.852 on SWE-bench Verified for a fifth of Fable 5. One of the best value-per-point models in the catalogue.

Previous-generation Opus. Still highly autonomous on long-horizon agentic work.

Fast and cheap for simple, high-volume work.

Alibaba's newest flagship — 2.4T-parameter MoE, natively multimodal, third in the creative-writing arena. Flat pricing across the whole 1M window, with no long-context surcharge.

Outstanding knowledge-task quality for the money — 0.885 MMLU-Pro at a fraction of frontier pricing.

The floor of the catalogue. For classification, routing and high-volume trivial chat.

Open-weight coding specialist with a very large output budget.

OpenAI's flagship. Leads Terminal-Bench 2 and tops the GPQA science leaderboard.

The balanced GPT-5.6 tier. Strong general and agentic performance at mid-range pricing.

Small, quick and very cheap, with the full million-token window.

Punches far above its price on maths — fourth in the arena, level with models costing several times more.

Google’s reasoning flagship, strong across science and long-context work.

Cheap, fast, million-token context. A good default for summarisation and extraction.

Frontier-adjacent quality at a tenth of frontier prices, and by far the most aggressive prompt-cache rate on the market — cached input costs under 1% of fresh.

Remarkable coding quality for the price — 0.79 SWE-bench at pennies per million tokens.

Open-weight flagship that effectively saturates scripted tool-use benchmarks (99.1% on τ²-bench).

Very cheap, and near-perfect on scripted tool workflows (98.8% τ²-bench).

Statistically level with Gemini 3.1 Pro on SWE-bench Verified at a fraction of the price. The clearest example of why routing pays.

Top-tier agentic coding, and unusually high prompt-cache hit rates in production.

Open-weight coding specialist tuned for long edit sessions.

Strong general-purpose reasoning with competitive coding scores.

European-hosted general model with solid coding performance.