Coding
Which model should write and fix code. Four answers — best value for money, best there is, best with open weights, cheapest that still works.
Which model should write and fix code. 113 measured, 108 of them for sale, and 64 of those stand on two or more of these boards — only those can be picked.
- Best value for moneyMiMo V2.6 ProXiaomi · 9th of 235 on Scicode · 3 boards$0.37→$0.74per Mtok in / out11 sellers
- Best frontierClaude Fable 5Anthropic · 1st of 29 on Surface evolver bench · 2 boards$10→$50per Mtok in / out15 sellers
- Best open sourceGLM 5.3 FlashZ.ai · 23rd of 135 on Webdev arena · 4 boards$0.022→$0.12per Mtok in / out25 sellers
- CheapestGPT-5OpenAI · 73rd of 202 on Terminalbench · 2 boards$0.04→$0.08per Mtok in / out16 sellers
Judged on
- Aider polyglot coding leaderboard
- LMArena · WebDev
- Scicode — Epoch AI
- Cursorbench — Epoch AI
- Frontiercode — Epoch AI
- Terminalbench — Epoch AI
- Webdev arena — Epoch AI
- Algotune — Epoch AI
- Surface evolver bench — Epoch AI
Capability is a model's mean percentile across the boards below, on each board's widest metric. Prices are quoted as you are billed them; where a model charges separately to read and to write, the two are ranked against each other at one part in to three parts out, which for this field puts the middle at $1.95 per million tokens. Best value for money is the most capable at or under that middle, and cheapest is the least expensive of those above the middle for capability — without that second floor the cheapest answer is reliably the worst model in the category. Figures read 2026-10-31.