MODEL VERSIONS ON THIS PAGE HAVE MOVED — CHECKED 8 OCTOBER 2026
Current line, with published API pricing per million tokens:
●
OpenAI — GPT-6 Sol (paid ChatGPT), GPT-6 Luna ($0.10 / $0.50), GPT-6.1 Sol ($2.00 / $10.00), GPT-6 Astra
●
Anthropic — Claude Opus 5.5, Sonnet 5.5 ($2.00 / $10.00), Haiku 5.5 ($0.10 / $0.50 below 100K), Fable 5.1
●
xAI — Grok 4.3 ($1.25 / $2.50), Grok 4.7 ($2.00 / $6.00)
●
Google — Gemini 3.5 Flash-Lite ($0.30 / $2.50), Gemini 3.8 Flash ($0.75 / $3.75 to 31 Dec), Gemini 4 Argon
The comparison below is kept because the reasoning still holds — how these tiers trade price against capability has not changed. The version numbers and prices have.
Current pricing table →
QUICK VERDICT — JULY 30, 2026
● SWE-bench Pro: Fable 5 — 80.4% #1 vs Laguna 59.4%
● Price: Laguna — $0.10/$0.20/M vs Fable 5 $10/$50/M (100× cheaper input)
● Terminal-Bench 2.1: Laguna — 70.2% vs Fable 5 83.8% in Claude Code
● Open weights: Laguna only — self-hostable on DGX Spark
● Regulatory risk: Laguna — none. Fable 5 — had export control suspension (resolved).
● Context: Both 1M tokens
Use Laguna S 2.1 for: High-volume coding workloads where the 59.4% SWE-bench Pro score is sufficient (roughly comparable to where GPT-4 was on SWE-bench in 2025), where self-hosting is valuable, and where 100× lower cost changes the economics of your use case fundamentally.
Use Fable 5 for: Production code changes where SWE-bench Pro 80.4% accuracy is specifically required and independently verified. If your engineering team has validated that the 21-point SWE-bench gap translates to meaningfully fewer bugs in your codebase — and it likely does — the $10/M premium is justified for critical production work.
Last updated July 30, 2026. Related: Laguna S 2.1 full review →