Benchmarks
More
OverviewBenchmaxxingCompareChartsCost vs CapabilityProviders per ModelProvider explorerGatewaysEU & SovereignJev-class modelsAbout
← All models

GPT-5.6 Sol (max)

★ featured
OpenAI · released 2026-07-09 · 10 offers

Output 60 tokens/sFirst token 57 sContext 1M tokens

Top 5 cheapest providers (Adjusted $/task)

The same list price can give a different adjusted $/task (caching, token efficiency) — click a price for its inputs.

Within the active global provider, residency and confidentiality filters.

#ProviderAdjusted $/task
1Amazon Bedrock
2Azure
3Azure AI Foundry
4AWS Bedrock
5T-Systems LLM Hub

Composite

5 of 7 inputs · 2 from the model family6 radar axes: DesignArena's two boards share one

95.3

AA CodingCoding Agent v1.4AA IntelligenceAA AgenticEpoch ECISoftware ECIDesignArenaAA Coding: percentile 98Coding Agent v1.4: percentile 86AA Intelligence: percentile 98Epoch ECI: percentile 90Software ECI: percentile 84

AA Coding 77.4Coding Agent v1.4 65.1AA Intelligence 47.1AA Agentic —Epoch ECI 161.8Software ECI 160.5DesignArena —/—

Radar: percentile among all models measured on each input; a gap means not measured.

Benchmark sheet

56 of 140 registered benchmark versions · bars show the percentile among all models measured on each benchmark.

Composite attachments (used in the score, not counted as exact benchmarks):
  • Epoch ECI · attached
  • Software ECI · attached
Compare this model →

Agentic

  • AA-AnalystAgent8147.5%

    AA-AnalystAgent published 2026-09-10 ↗ · Published board — Tests analyst tasks using agentic Python execution across fourteen domains.

  • AA-Briefcase871,475 Elo

    AA-Briefcase published 2026-09-10 ↗ · Published board — Tests multi-week professional knowledge-work projects with linked tasks and large source collections.

  • GDPval-AA v2921,624 Elo

    GDPval-AA v2 v2 ↗ · Published board — Tests professional knowledge-work deliverables across occupations using AA's Stirrup harness.

  • Harvey LAB-AA6287.2%

    Harvey LAB-AA published 2026-09-10 ↗ · Published board — Tests legal-work deliverables across practice areas on Harvey's private task set. Artificial Analysis' run, graded by one LLM judge against task rubrics — not the same run or scale as Vals AI's HLAB row.

  • ITBench-AA10056.2%

    ITBench-AA published 2026-09-10 ↗ · Published board — Tests root-cause diagnosis from offline Kubernetes incident snapshots.

  • Terminal-Bench Hard (AA)10065.9%

    Terminal-Bench Hard (AA) pinned revision 74221fb ↗ · Published board — Tests a pinned 44-task hard subset of terminal-based work using Terminus 2.

  • Terminal-Bench v2.1 (AA)9588.0%

    Terminal-Bench v2.1 (AA) v2.1 ↗ · Published board — Tests terminal-based work on the 89-task verified refresh using Terminus 2.

  • Terminal-Bench v4.0 (AA)9039.9%

    Terminal-Bench v4.0 (AA) v4.0 ↗ · Published board — Tests terminal-based work on the 66-task release using mini-SWE-agent v2.4.6.

  • ApprenticeBench CUA (NeoCognition)8226.0%

    ApprenticeBench CUA (NeoCognition) published 2026-09-14 ↗ · Codex — A computer-use agent operates the Odoo ERP through its screens across 100 sequential accounts-payable tasks with diminishing mentoring.

  • ApprenticeBench API (NeoCognition)7530.0%

    ApprenticeBench API (NeoCognition) published 2026-09-14 ↗ · Codex — An agent completes the same 100 sequential accounts-payable tasks in the Odoo ERP with diminishing mentoring, using dedicated application API calls instead of the screen interface.

  • Vals Index v2 (Vals AI)8363.7%

    Vals Index v2 (Vals AI) v2 ↗ · Published board — GDP-weighted average of agentic model performance across finance, coding and legal tasks.

  • Finance Agent v2 (Vals Index v2)4753.8%

    Finance Agent v2 (Vals Index v2) v2 ↗ · Published board — Multi-step financial reasoning tasks.

  • Terminal-Bench 2.1 (Vals Index v2)9785.8%

    Terminal-Bench 2.1 (Vals Index v2) v2 ↗ · Published board — Command-line interface problem solving, as run by Vals AI.

  • HLAB — Harvey's Legal Agent Benchmark (Vals Index v2)232.5%

    HLAB — Harvey's Legal Agent Benchmark (Vals Index v2) v2 ↗ · Published board — Long-horizon legal work product creation. Vals AI's run of Harvey's Legal Agent Benchmark, reported as accuracy — not comparable with Artificial Analysis' Harvey LAB-AA row.

  • OSWorld 2.0, August 2026 task release (XLANG Lab) v2026.08.08no percentile27.3%

    OSWorld 2.0, August 2026 task release (XLANG Lab) v2026.08.08 ↗ · batch tool — A computer-use agent completes 108 long-horizon, real-world workflows across 31 self-hosted websites and desktop applications, each taking a person about 1.6 hours.

  • EBR-bench (Earthborne Rangers, Epoch AI)7744.8%

    EBR-bench (Earthborne Rangers, Epoch AI) published 2026-09-18 ↗ · Published board — Learning-capability test: models repeatedly play the obscure campaign board game Earthborne Rangers with note-taking, and the score measures whether results improve across playthroughs.

  • PostTrainBench v1.1no percentile36.2%

    PostTrainBench v1.1 ↗ · Codex CLI — Agents retrain four small base models (gemma-3-4b-pt, SmolLM3-3B-Base, Qwen3-1.7B-Base, Qwen3-4B-Base) for each of 7 benchmark families; the leaderboard value is the weighted mean over base models × benchmarks, aggregated over the 2–3 listed runs per agent.

Coding

  • Artificial Analysis Coding Agent Index v1.49565.1%

    Artificial Analysis Coding Agent Index v1.4 v1.4 ↗ · Codex — The retained AA Coding Agent Index measures coding-agent systems using the earlier three-component implementation.

  • Artificial Analysis Coding Agent Index v1.56754.6%

    Artificial Analysis Coding Agent Index v1.5 v1.5 ↗ · Codex — Measures coding-agent systems on DeepSWE v1.1, Terminal-Bench 4.0 and SWE-Atlas-QnA.

  • SciCode (AA subproblems) v1.0.19157.1%

    SciCode (AA subproblems) v1.0.1 ↗ · Published board — Tests scientific Python programming with scientist-annotated background information.

  • CursorBench 4.0 (Cursor) v4.0no percentile41.7%

    CursorBench 4.0 (Cursor) v4.0 ↗ · Published board — Agent evaluation on ambiguous, multi-file tasks drawn from real Cursor sessions.

  • Vibe Code Bench (Vals Index v2)7480.5%

    Vibe Code Bench (Vals Index v2) v2 ↗ · Published board — End-to-end app-building tasks.

  • Code Migration (Vals Index v2 subset)8852.4%

    Code Migration (Vals Index v2 subset) v2 ↗ · Published board — Porting projects to another language, including COBOL modernization.

  • FrontierCode 1.1 Main (Cognition) v1.1no percentile47.5%

    FrontierCode 1.1 Main (Cognition) v1.1 ↗ · codex — Whether a maintainer would merge the agent's pull request, on tasks crafted by open-source maintainers and graded with tests, rubrics and verifiers.

  • DeepSWE (Datacurve, via Epoch AI)8772.7%

    DeepSWE (Datacurve, via Epoch AI) published 2026-09-15 ↗ · mini-swe-agent — Pass rate of coding agents on original, long-horizon software engineering tasks, run by Datacurve with the mini-swe-agent harness.

  • FrontierSWE v28632.2%

    FrontierSWE v2 v2 ↗ · proximus — 34 hand-written real-world software tasks: each model attempts every task in its own CLI agent harness for 5 trials per task under a 20-hour budget, scored by the site’s own review protocol; the headline value is mean@5 in percent with best@5/worst@5 bounds.

  • AA Coding Index9877.4

    AA Coding Index published 2026-09-19 ↗ · Published board — Artificial Analysis publishes this board without a version number; we keep the date each result was retained.

Efficiency

Instruction-following

Knowledge

Long-context

  • GDP.pdf (AA)9527.2%

    GDP.pdf (AA) published 2026-09-10 ↗ · Published board — Tests professional reasoning over long PDFs with AA document preparation and grading.

  • AA-LCR v1.19984.0%

    AA-LCR v1.1 v1.1 ↗ · Published board — Tests reasoning across multiple long documents with corrected answer keys and grading.

  • MLCR-AA8026.1%

    MLCR-AA published 2026-09-10 ↗ · Published board — Tests medical-record synthesis and reasoning across long, fragmented case documents.

Math

  • ArXivMath 06/2026 (MathArena) v2026-068088.5%

    ArXivMath 06/2026 (MathArena) v2026-06 ↗ · Published board — Research-level math problems with a checkable final answer, drawn from arXiv papers submitted in June 2026, so they postdate most training data.

  • BrokenArXiv 06/2026 (MathArena) v2026-066067.3%

    BrokenArXiv 06/2026 (MathArena) v2026-06 ↗ · Published board — Plausible but false proof statements taken from June 2026 arXiv papers; a model scores by refusing to prove them and saying the statement is false as written.

  • FrontierMath Tiers 1–3 v2 (Epoch AI)9489.1%

    FrontierMath Tiers 1–3 v2 (Epoch AI) v2 ↗ · Published board — Unpublished, expert-written mathematics problems from undergraduate to research level with automatically checkable answers, run by Epoch AI on its private v2 set.

  • FrontierMath Tier 4 v2 (Epoch AI)7882.9%

    FrontierMath Tier 4 v2 (Epoch AI) v2 ↗ · Published board — The hardest, research-level tier of Epoch AI's unpublished FrontierMath problems, run by Epoch AI on its private v2 set.

Reasoning

  • Chess Puzzles (Epoch AI)9655.0%

    Chess Puzzles (Epoch AI) published 2026-09-18 ↗ · Published board — Best-move selection on 100 novel chess positions generated programmatically by Epoch AI, each with a single Stockfish-verified best move; probes spatial reasoning and planning.

  • Mystery Game Puzzles (Epoch AI)9358.0%

    Mystery Game Puzzles (Epoch AI) published 2026-09-18 ↗ · Published board — Best-move selection on 100 mid-game positions of a well-known game whose identity Epoch deliberately keeps undisclosed, generated programmatically like Chess Puzzles.

  • AA Intelligence Index9847.1

    AA Intelligence Index published 2026-09-19 ↗ · Published board — Artificial Analysis publishes this board without a version number; we keep the date each result was retained.

Safety/Alignment

Science

Tool-use

Vision

Unusual results

3 threshold-crossing signals flagged · 41 eligible benchmark families

How flags are calculated

Heuristic screen, not statistical significance: benchmark families are correlated and source uncertainty is unknown. Peer evidence needs ≥ 20 independently measured matched configurations from ≥ 10 distinct model families. A flag needs a directed population z-score of magnitude ≥ 1.5 and a gap of ≥ 1.5 from the leave-one-benchmark-family-out mean z in the same direction, over ≥ 5 other benchmark families.

  • unusually strong CritPt (AA) published 2026-09-10
    Why
    Observed: 0.32286 fraction
    Peer mean 0.03928 · peer sd 0.07596 · n 522 · families 369
    Directed z 3.733 · baseline z 1.119 · gap 2.615 · profile n 40
    0.32286 fractionmeasuredobserved 2026-09-10artificialanalysis.ai ↗
    Evidence
    Axis: CritPt (AA) · published 2026-09-10 · Published board
    Exact value: 0.322857142857143 fraction
    Observed: 2026-09-10T21:47:16.627Z · publication date: not recorded
    Observation id: aa:d93edfe8-bf35-49ad-b56e-b18116142a1c:critpt
  • unusually strong Terminal-Bench Hard (AA)
    Why
    Observed: 0.65909 fraction
    Peer mean 0.18472 · peer sd 0.16951 · n 432 · families 315
    Directed z 2.798 · baseline z 1.142 · gap 1.656 · profile n 40
    0.65909 fractionmeasuredobserved 2026-09-10artificialanalysis.ai ↗
    Evidence
    Axis: Terminal-Bench Hard (AA) · pinned revision 74221fb · Published board
    Exact value: 0.659090909090909 fraction
    Observed: 2026-09-10T21:47:16.627Z · publication date: not recorded
    Observation id: aa:d93edfe8-bf35-49ad-b56e-b18116142a1c:terminalbenchHard
  • unusually strong AA Intelligence Index published 2026-09-19
    Why
    Observed: 47.1 points
    Peer mean 15.94317 · peer sd 11.48893 · n 644 · families 483
    Directed z 2.712 · baseline z 1.144 · gap 1.568 · profile n 40
    47.1 pointsmeasuredobserved 2026-09-19artificialanalysis.ai ↗
    Evidence
    Axis: AA Intelligence Index · published 2026-09-19 · Published board
    Exact value: 47.1 points
    Observed: 2026-09-19 · publication date: not recorded
    Observation id: legacy:aa_intelligence_index:gpt-5.6-sol::max
    Protocol: inspect · file data/raw/artificialanalysis.json
Missing coverage · 87 benchmark versions

No result does not mean a zero, or that the model was never tested. Collection failures and disputed versions retain their distinct status.

Variants / reasoning settings

Artificial Analysis snapshot 2026-09-19 · Data: Artificial Analysis · Data: Epoch AI (CC BY)

VariantAA CodingAA Intelligence
GPT-5.6 Sol (xhigh)78.344.1
GPT-5.6 Sol (max)77.447.1
GPT-5.6 Sol (high)77.242.5
GPT-5.6 Sol (medium)76.339.5
GPT-5.6 Sol (low)69.733.8
GPT-5.6 Sol (Non-reasoning)65.128.3

SUBSCRIPTION PLAN

GitHub Copilot

Current usage-based billing · model token cost is converted to AI Credits at 1 credit = $0.01.

Input $4.00 / 1M
Cached input $0.400 / 1M
Cache write $5.00 / 1M
Output $20.00 / 1M
Status GA
Token offers by platform · 8 offers (Adjusted $/task)

Click any underlined price to see how it is estimated and where each input comes from. How we calculate adjusted cost.

8 offers within the active global filters; “—” means the catalog is active but no public token price is available.

OpenRouter (4)

Amazon Bedrock
Azure
Azure
Azure

Azure AI Foundry (2)

Azure AI Foundry
Azure AI Foundry

AWS Bedrock (1)

AWS Bedrock

T-Systems LLM Hub (1)

T-Systems LLM Hub
GPT-5.6 Sol (max) — benchmarks & cost | Benchmark Heaven