THE SCOREBOARD FOR WHAT'S NEXTAI + QUANTUM, SCORED. COMPARED. EXPLAINED.
AIQUANTUMSCORE.

AI INTELLIGENCE / THE SCOREBOARD

Which AI is worth your time?

Compare performance, API prices and the evidence behind every score.

41Models tracked
16With overall scores
285Benchmark records
69Cited sources
Latest model verification2026-10-16Check source health →
Awaiting scoresBroader evidence required for a Quantum Score
Gemini 4 Argon ↗Google · UnscoredMuse Wonder ↗Meta · UnscoredMuse Wonder Full ↗Meta · Unscoredgpt-4.1 ↗OpenAI · Unscored

LIVEBENCH 2.0 / RELEASE 20261025 / CHECKED 2026-11-07

Independent coding benchmark

Tracked models in this release, ordered by published coding result. Equal results are ties.

ModelCoding %Reasoning %Long agentic %
Claude Fable 5.197.791.777.9
Claude Opus 5.596.392.272.8
GPT-6.1 Sol9692.668.7
Claude Sonnet 5.594.791.663.4
GPT-6 Astra93.792.777.5
Gemini 4 Argon89.795.161.2
Grok 4.689.191.228
GPT-5.6 Terra86.290.633.1
GPT-6 Luna84.581.830.9
Gemini 3.8 Flash82.289.316.5

LiveBench release data ↗ · Published benchmark percentages. Effort is unspecified in the release table. These results cover the named tasks and use a separate scale from Quantum Scores. Evaluation dates are unpublished.

PERFORMANCE + PRICE

Qualified Quantum Scores

All rankings →
#ModelQuantum ScoreCodingReasoningREVAL™
ⓘREVAL™ Research Evaluation Index — our proprietary 0–100 benchmark measuring real-world AI research performance. Methodology →
SpeedValueAPI $ / 1M tokens7d
01Claude Opus 5.5Anthropic88QUANTUM SCORE9491Not tested90 tok/sMax · 2026-10-0152In $4.00 / Out $20.00 ↗Verified 2026-09-29→
02Claude Sonnet 5.5Anthropic84QUANTUM SCORE9591Not testedNot measured62In $2.00 / Out $10.00 ↗Verified 2026-09-29→
03GPT-6 AstraOpenAI74QUANTUM SCORE7494Not tested52 tok/sMax · 2026-10-0127In $10.00 / Out $50.00 ↗Verified 2026-10-19→
04Claude Fable 5.1Anthropic68QUANTUM SCORE7086Not testedNot measured25In $10.00 / Out $50.00 ↗Verified 2026-09-29→
05Gemini 3.7 FlashGoogle62QUANTUM SCORE4791Not testedNot measured56In $0.75 / Out $3.75 ↗Verified 2026-09-29→
06DeepSeek V4.1 FlashDeepSeek60QUANTUM SCORE4692Not testedNot measured58In $0.30 / Out $1.20 ↗Verified 2026-09-29→
07Gemini 3.8 FlashGoogle57QUANTUM SCORE3261Not testedNot measured54In $0.75 / Out $3.75 ↗Verified 2026-09-29→
08GPT-5.6 TerraOpenAI54QUANTUM SCORE4967Not tested93.6 tok/sMax · 2026-10-0140In $2.00 / Out $12.00 ↗Verified 2026-10-19→
09Gemini 3.5 FlashGoogle53QUANTUM SCORE3275Not testedNot measured43In $1.50 / Out $9.00 ↗Verified 2026-09-29→
10GPT-6 SolOpenAI52QUANTUM SCORE4767Not testedNot measured39In $2.00 / Out $10.00 ↗Verified 2026-10-19→

Scores are 0–100, subject to evidence coverage. API prices are USD per million tokens; source links and verification dates are shown with each price. Missing evidence stays unscored. How we score →

RANK HISTORY

Biggest movers

Timeline →
No verified rank movement yet. Daily snapshots are building the history; unchanged rankings do not create a new move.

PERFORMANCE PER DOLLAR

Value shortlist

Explore value →

THE CONTEXT

News & analysis

All news →
AI

World Labs opens Atlas world-model tools and APIs

Atlas moves beyond a model demonstration into public tools for reconstructing, shaping and exporting explorable 3D scenes, with developer endpoints intended for creative, simulation and robotics workflows.

News brief · 3 sourcesRead →

START HERE

What brought you here?

Pick a path. You can explore the tools without creating an account.

GO DEEPER

Evidence, tools & the frontier