Skip to main content

AI Model Index

Pricing sourced from Artificial Analysis. Intelligence Index, accuracy, recall, speed and Coding Agent Index results from Artificial Analysis.
Last updated: November 19, 2026

26 models

Models with a verified Coding Agent Index v2.1 result: an exact tested harness, model variant and effort. A model-level AA Coding Index table is shown here only when Artificial Analysis publishes one.

Coding Agent Index

Colored cells grade each metric in quartiles relative to the full index for this table — top 25% green, bottom 25% red; shade intensity reflects position within the quartile. Search narrows the rows shown but does not change colors. Sorted by Coding Agent Index (highest first).

Claude Fable 5.2 (Xhigh, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
62
Omni Acc
78.0%
Recall
82.7%

Speed

Tok/s
87 tok/s
Time/task
31.2 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$14.83
Claude Fable 5.2 (High, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
62
Omni Acc
77.2%
Recall
83.3%

Speed

Tok/s
79 tok/s
Time/task
20.7 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$9.08
Claude Opus 5.5 (Max, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
59
Omni Acc
65.5%
Recall
84.7%

Speed

Tok/s
97 tok/s
Time/task
29.8 min

Price

Input / Output ($/1M)
$4.00 / $20.00
Cost
$10.04
Claude Sonnet 5.5 (Max, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
55
Omni Acc
51.5%
Recall
82.7%

Speed

Tok/s
130 tok/s
Time/task
36.5 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$10.02
Claude Fable 5.2 (Max, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
65
Omni Acc
79.1%
Recall
83.3%

Speed

Tok/s
88 tok/s
Time/task
51.8 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$26.54
GPT-6.2 Astra (Max)

OpenAI

Intelligence & accuracy

Intel
62
Omni Acc
68.6%
Recall
—
Agent config
53.5Codex · Max

Speed

Tok/s
68 tok/s
Time/task
10.1 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$4.00
GPT-6.2 Astra (Xhigh)

OpenAI

Intelligence & accuracy

Intel
62
Omni Acc
67.2%
Recall
—
Agent config
53.1Codex · Xhigh

Speed

Tok/s
48 tok/s
Time/task
7.7 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$2.52
GPT-6 Astra (Max)

OpenAI

Intelligence & accuracy

Intel
57
Omni Acc
61.6%
Recall
80.7%
Agent config
53.1Codex · Max

Speed

Tok/s
47 tok/s
Time/task
18.3 min

Price

Input / Output ($/1M)
$10.00 / $50.00
Cost
$6.24
Claude Fable 5.2 (Medium, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
56
Omni Acc
77.0%
Recall
80.0%

Speed

Tok/s
75 tok/s
Time/task
9.5 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$3.80
Claude Fable 5.1 (Max, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
49
Omni Acc
68.4%
Recall
85.3%

Speed

Tok/s
66 tok/s
Time/task
29.8 min

Price

Input / Output ($/1M)
$10.00 / $50.00
Cost
$12.99
GPT-6.2 Astra (High)

OpenAI

Intelligence & accuracy

Intel
60
Omni Acc
66.3%
Recall
—
Agent config
52.3Codex · High

Speed

Tok/s
42 tok/s
Time/task
7.1 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$2.07
GPT-6.1 Sol (Max)

OpenAI

Intelligence & accuracy

Intel
55
Omni Acc
61.8%
Recall
83.0%
Agent config
52.2Codex · Max

Speed

Tok/s
71 tok/s
Time/task
15.8 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$1.27
GPT-6.2 Sol (Xhigh)

OpenAI

Intelligence & accuracy

Intel
56
Omni Acc
60.7%
Recall
—
Agent config
49.8Codex · Xhigh

Speed

Tok/s
91 tok/s
Time/task
4.5 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$0.65
GPT-6.2 Astra (Medium)

OpenAI

Intelligence & accuracy

Intel
57
Omni Acc
64.2%
Recall
—

Speed

Tok/s
41 tok/s
Time/task
5.4 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$1.62
GPT-6.2 Sol (Max)

OpenAI

Intelligence & accuracy

Intel
57
Omni Acc
61.6%
Recall
—
Agent config
48.6Codex · Max

Speed

Tok/s
91 tok/s
Time/task
7.4 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$1.01
GPT-6.1 Terra (Max)

OpenAI

Intelligence & accuracy

Intel
51
Omni Acc
47.6%
Recall
82.7%
Agent config
48.6Codex · Max

Speed

Tok/s
80 tok/s
Time/task
12.4 min

Price

Input / Output ($/1M)
$1.00 / $5.00
Cost
$0.96
Claude Fable 5.2 (Low, Default Fallback)

Anthropic

Intelligence & accuracy

Intel
54
Omni Acc
76.3%
Recall
80.0%

Speed

Tok/s
79 tok/s
Time/task
7.6 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$3.10
GPT-6.2 Sol (High)

OpenAI

Intelligence & accuracy

Intel
55
Omni Acc
61.4%
Recall
—
Agent config
47.5Codex · High

Speed

Tok/s
91 tok/s
Time/task
3.7 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$0.56
GPT-6.2 Sol (Medium)

OpenAI

Intelligence & accuracy

Intel
53
Omni Acc
59.2%
Recall
—

Speed

Tok/s
83 tok/s
Time/task
2.7 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$0.41
Grok 4.7 (Xhigh)

SpaceXAI

Intelligence & accuracy

Intel
36
Omni Acc
42.7%
Recall
76.7%

Speed

Tok/s
70 tok/s
Time/task
25.0 min

Price

Input / Output ($/1M)
$2.00 / $6.00
Cost
$7.77
GLM-5.3 (Max)

Z AI

Intelligence & accuracy

Intel
32
Omni Acc
32.3%
Recall
79.7%

Speed

Tok/s
68 tok/s
Time/task
29.3 min

Price

Input / Output ($/1M)
$1.40 / $4.40
Cost
$5.01
GPT-6.2 Sol (Low)

OpenAI

Intelligence & accuracy

Intel
41
Omni Acc
54.6%
Recall
—
Agent config
37.2Codex · Low

Speed

Tok/s
88 tok/s
Time/task
1.1 min

Price

Input / Output ($/1M)
$2.00 / $10.00
Cost
$0.18
DeepSeek V4.1 Flash 1017 (Max)

DeepSeek

Intelligence & accuracy

Intel
33
Omni Acc
39.6%
Recall
82.7%

Speed

Tok/s
161 tok/s
Time/task
14.4 min

Price

Input / Output ($/1M)
$0.30 / $1.20
Cost
$0.60
GPT-6.2 Astra (Low)

OpenAI

Intelligence & accuracy

Intel
46
Omni Acc
60.6%
Recall
—
Agent config
35.6Codex · Low

Speed

Tok/s
41 tok/s
Time/task
2.1 min

Price

Input / Output ($/1M)
$8.00 / $40.00
Cost
$0.66
Claude Haiku 5.5 (Max)

Anthropic

Intelligence & accuracy

Intel
37
Omni Acc
34.7%
Recall
82.7%

Speed

Tok/s
227 tok/s
Time/task
16.6 min

Price

Input / Output ($/1M)
$0.10 / $0.50
Cost
$0.41
GPT-6 Luna (Max)

OpenAI

Intelligence & accuracy

Intel
28
Omni Acc
39.6%
Recall
83.3%
Agent config
25.9Codex · Max

Speed

Tok/s
119 tok/s
Time/task
10.6 min

Price

Input / Output ($/1M)
$0.10 / $0.50
Cost
$0.15

Notes

  • Intelligence Index = Artificial Analysis Intelligence Index v4.3.2 (composite of 10 evaluations: AA-Briefcase v1.1, GDPval-AA v2.1, AutomationBench-AA, Terminal-Bench 4.0, SciCode, Humanity’s Last Exam, GDP.pdf, CritPt, AA-Omniscience, and AA-LCR v1.1). Higher is better. Scores reflect max-effort / adaptive reasoning mode where applicable.
  • AA-Omni Accuracy = AA-Omniscience Accuracy (%). Higher is better. Proportion of correctly answered questions out of all questions. “—” means AA has not published a value for the model. See artificialanalysis.ai/evaluations/omniscience.
  • Recall = Artificial Analysis Long Context Reasoning (AA-LCR) recall accuracy (%). Higher is better. “—” means AA has not measured the model. See artificialanalysis.ai/evaluations/artificial-analysis-long-context-reasoning.
  • Cost per Task = Weighted average pay-per-token API cost (USD) per Artificial Analysis Intelligence Index task. Lower is better. Reflects input, cache, reasoning, and output token pricing across index evaluations.
  • AA Coding Index = The model-level coding score supplied in Artificial Analysis’ language-model dataset (`artificial_analysis_coding_index`). See Artificial Analysis Data API documentation. This is distinct from the separate Coding Agent Index leaderboard, which reports results for specific agent + model variant + effort configurations. “—” means AA has no model-level score for that model in the dataset. It is shown only when AA publishes it, and AA currently publishes no model-level AA Coding Index, so this column and table are hidden.
  • Time/task = Artificial Analysis Intelligence Index output tokens per task ÷ the model’s median output speed. Where AA publishes no direct per-task time, we compute it from those two AA values. “—” means AA has published neither.
  • Agent Config = AA Coding Agent Index score for one exact tested harness + model variant + effort configuration. Results are specific to the published, versioned index and agent version; a score does not generalize to other agents or model efforts. “—” means no exact verified local binding. See AA Coding Agent Index leaderboard and benchmark methodology.
  • Data last fetched from Artificial Analysis — Model sync (pricing, Intelligence Index, cost and speed): November 19, 2026; Coding Agent Index v2.1: November 19, 2026; Omni Acc: November 19, 2026; Recall: November 19, 2026; Time/task: November 19, 2026.
  • The sync keeps separate AA source pools: top model-level Coding Index rows for Stable and top Intelligence Index rows without that score from trusted providers for Experiment; curated Qwen open-weight rows form Local. At render time Stable is defined by the models with a verified Coding Agent Index result, plus any model with an AA Coding Index score when AA publishes one; the category count includes a model with both scores once. Each benchmark table appears only when the pool has data for it.
  • Exact Agent Config results are refreshed independently with `pnpm fetch:aa-coding-agent-index`; only explicit local model/result ID bindings that match the expected harness, model variant, effort, and index version are shown. Null means no verified exact binding, not that the model has no agent result.
  • Coding Agent Index results never carry across index versions; a binding that does not match the published version shows “—”.
  • Coding Agent Index and AA Coding Index are independent measurements: each benchmark table sorts and grades only its own score. Exact agent results require a verified harness, model variant, effort, and index version. Gemini 4 Argon’s agent result has no published effort, so it is not attached to the High model row and remains on Experiment.
  • Accuracy rates are sourced from AA-Omniscience via SSR scrape snapshot (`aa-omniscience-snapshot.json`, refreshed with `pnpm fetch:aa-omniscience`). Not included in the free AA sync API — see memory/development/ai-model-index-omniscience-scrape.md.
  • Recall is sourced from the public AA-LCR SSR snapshot (`aa-lcr-snapshot.json`, refreshed with `pnpm fetch:aa-lcr`) because the free AA sync API omits per-benchmark fields. Null means AA has not measured the model.
  • Pricing reflects standard (non-cached) token rates from Artificial Analysis.
  • OpenAI cut GPT-5.6 Luna API pricing 80% and Terra 20% on July 30, 2026 (Luna $0.20/$1.20, Terra $2/$12 per 1M; Sol unchanged).
  • Gemini 4 Argon initially has limited access through Google Fairwind. It has no exact Coding Agent Index binding because its AA result has no published effort, so it appears on Experiment.
  • Claude Opus 4.7 and Claude Sonnet 4.6 both support a 1M token context window with 128K max output tokens.
  • Claude Fable 5.1 (Max) is on Stable through its verified Coding Agent Index result; its other efforts have no exact binding and stay on Experiment.
  • Claude Fable 5 and Claude Mythos 5 (Fable 5's unrestricted-cyber/bio counterpart) launched June 9, 2026. Access was suspended June 12–July 1, 2026 for U.S. export-control compliance; current pricing/scores reflect the restored, generally available model.
  • GPT-5.6 ships as three durable capability tiers: Sol (flagship, AA-Omni index 21.7 / accuracy 58.5%), Terra (balanced, -0.2 / 45.9%), and Luna (fastest/cheapest, -11.2 / 41.5%).
  • Grok 4.5 (high) is xAI's first release since its SpaceXAI rebrand and IPO; it's notably cost-efficient per completed task with an AA-Omniscience accuracy of 52.1% (index 26.4).
  • Open-weight models (self-hostable): Kimi K3 (weights releasing July 27, 2026), Kimi K2.7 Code, Kimi K2.6, MiMo-V2.5-Pro, GLM 5.3, GLM 5.3 Flash, DeepSeek V4 Pro, DeepSeek V4 Flash, Hy3-preview, Nemotron 3 Super.
  • DeepSeek V4 Pro and V4 Flash OpenRouter pricing shown; first-party DeepSeek API pricing differs ($1.74/$3.48 for Pro, $0.14/$0.28 for Flash).
  • Nemotron 3 Super is also available free-tier on OpenRouter (rate-limited).

Changelog

DateChange
2026-11-18Artificial Analysis stopped publishing the model-level AA Coding Index. Stable is now defined at render time by verified Coding Agent Index v2.1 results, and the AA Coding Index table and column are hidden until AA republishes.
2026-11-18Added and bound the newest Anthropic, OpenAI, DeepSeek and GLM models with all their effort variants, and refreshed Recall, Omni Acc and Time/task for them. The notes now show each source’s fetch date.
2026-10-08Added Claude Haiku 5.5 Coding Agent Index results (Max, Xhigh, High, Medium, Low) from the official leaderboard. Haiku 5.5 now appears on the Stable tab. AA Coding Index is not published for Haiku 5.5.
2026-10-08Refreshed AA data and added Claude Haiku 5.5 (Max, Xhigh, High, Medium, Low) to the Experiment tab with Intelligence Index, pricing, speed, and time-per-task scores. Recall and Omni Acc are shown where AA publishes them. Haiku 5.5 has no model-level AA Coding Index score, so it is not in Stable.
2026-10-05Stable now includes models with either a model-level AA Coding Index score or an exact verified Coding Agent Index result. The primary Coding Agent Index and secondary AA Coding Index comparisons remain independent.
2026-10-05Refreshed benchmark measurements and added verified GPT-6.1 Sol, GPT-6 Sol, GPT-6 Luna, Sonnet 5.5, and Grok 4.7 agent configurations to Experiment.
2026-10-01Added Gemini 4 Argon to Experiment and refreshed curated frontier comparisons across Stable and Experiment from Artificial Analysis.
2026-09-30Added GPT-6.1 Sol (max, xhigh, high, medium, low) to the Experiment comparison from Artificial Analysis. Model-level AA Coding Index scores are not available yet.
2026-09-16Expanded the Stable and Experiment tabs from the top 25 to the top 50 models each via the model sync limits; refreshed model lists and benchmark snapshots.
2026-09-16Refreshed Stable, Experiment, and Local model lists and benchmark snapshots from Artificial Analysis; added DeepSeek V4.1 Flash and Claude Sonnet 5 effort variants (xhigh, high, medium, low) to the Experiment tab.
2026-09-04Added AA-LCR Recall as a separate benchmark. Configured Grok fast variants use the base Grok 4 recall while retaining their own source pricing and identity.
2026-09-03Added Grok 4 Fast (Reasoning and Non-reasoning) as separate Experiment rows with duplicated Grok 4 benchmark scores and Artificial Analysis speed-specific token pricing.
2026-09-03Refreshed Stable, Experiment, and Local model data from Artificial Analysis; added Claude Fable 5.1 effort variants to Stable and Claude 4.1 Opus (Reasoning) to Experiment.
2026-09-02Added Claude Fable 5.1 (Anthropic) to Experiment tab — AA Intelligence Index 66 (#1). Pinned Claude Fable 5 on Stable tab; refreshed Fable 5 benchmarks.
2026-09-01Refreshed pricing and cost-per-task metrics from Artificial Analysis for all Stable, Experiment, and Local tab models.
2026-08-27Added GLM 5.3 Flash on Stable tab — AA Intelligence Index 58, AA Coding Index 71.5.
2026-08-20Added GLM 5.3 (max) benchmarks on Stable tab — AA Intelligence Index 60, AA Coding Index 74.8.
2026-08-20Refreshed DeepSeek V4 Pro 0813 benchmarks on Stable tab — latest GA checkpoint (deepseek-v4-pro).
2026-08-18Added Local tab with curated Qwen 3.6/3.8 open-weight models.
2026-08-15Refreshed AA time-per-task snapshot — all tracked models now have Time/task metrics including Luna effort variants; fixed prior timeout failures.
2026-08-15Added GPT-5.6 Luna thinking-effort variants (xhigh, high, medium) to Stable tab alongside existing Luna (max).
2026-08-14Restored DeepSeek V4 Pro 0813 on Stable tab — AA slug renamed to deepseek-v4-pro.
2026-08-15Models with announced pricing changes show a warning icon beside the input/output price — tap or click for details.
2026-08-14Replaced sort dropdown with Coding / Cost / Speed buttons — single-click primary dimension for model list ordering.
2026-08-14Added Tok/s and Time/task columns to desktop table; card and table now share a single field registry so metrics stay in sync.
2026-08-14Added AA time-per-task scores via SSR snapshot — median output speed and intelligence-index output tokens per task scraped from Artificial Analysis model pages.
2026-08-14Refreshed AA-Omniscience accuracy snapshot for all 50 tracked models — 49/50 populated. Claude 4.1 Opus (Reasoning) remains unbenchmarked on AA (omniscience null; deprecated).
2026-08-14Pinned GLM 5.3 (Z.ai) for Stable tab — awaiting Artificial Analysis benchmark publication.
2026-08-13Added DeepSeek V4 Pro 0813 benchmarks (intelligence 53, coding index 68.8) on Stable tab.
2026-08-13Added Grok 4.6 (high) to Stable benchmarks — AA Intelligence Index 61, AA Coding Index 76.8.
2026-08-03Added Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) to Stable benchmarks — AA Coding Index 63.
2026-08-01Added output-price sort to model table. Pinned Claude Haiku 4.5 (Reasoning) and GPT-5.6 Luna on Stable tab.
2026-08-01Synced GPT-5.6 Luna and Terra price reductions (Luna $0.20/$1.20, Terra $2/$12 per 1M; Sol unchanged at $5/$30) following OpenAI July 30, 2026 API price cut.
2026-07-31Updated DeepSeek V4 Flash 0731 benchmarks (intelligence 50, coding index 69.1) on Stable tab.
2026-07-25Wired AA-Omniscience hallucination rates for all 50 tracked models (stable + experiment) from SSR scrape snapshot.
2026-07-25Fixed benchmark column to show AA-Omniscience Accuracy instead of conditional hallucination rate — values now match AA's prominently displayed metric (e.g. GPT-5.6 Sol ~58.5%, Claude Opus 5 ~54.2%).
2026-07-24Synced Claude Opus 5 (max, xhigh, high, medium) from Artificial Analysis — Opus 5 (max) leads Intelligence Index at 61 and Coding Index at 78.
2026-07-20Added Qwen3.7 Max (Alibaba) to Stable benchmarks. Qwen 3.8 is not yet published on Artificial Analysis — pinned model IDs will pick it up automatically on the next sync once AA adds it.
2026-07-19Added Cost/Task column with relative color scale on Stable and Experiment model cards and tables.
2026-07-18Switched to AA-synced stable/experiment model lists (sync:model-index) instead of manual models.json.
2026-07-17Added Kimi K2.7 Code (Moonshot AI).
2026-07-17Replaced Agentic Index and AA-Omni index with Intelligence Index, hallucination rate, and AA Coding Index columns.
2026-07-17Added Kimi K3 (Moonshot AI).
2026-07-10Added Grok 4.5 (xAI).
2026-07-10Added Claude Fable 5 (Anthropic).
2026-07-10Added GPT-5.6 Sol, Terra, and Luna (OpenAI).
2026-05-17Added MiniMax M2.7 (MiniMax).
2026-05-17Added Hy3-preview (Tencent).
2026-05-17Filled in AA-Omni for Gemini 3 Flash (12), GPT-5.5 xhigh (18, corrected from 20), Grok 4.3 high (18).
2026-05-17Added GPT-5.5 xhigh (OpenAI).
2026-05-17Added Grok 4.3 high (xAI).
2026-05-17Added Gemini 3 Flash Preview (Google).
2026-05-17Filled in AA-Omni scores for MiMo-V2.5-Pro (4), Qwen3.6 Plus (3), Nemotron 3 Super (-42).
2026-05-17Added MiMo-V2.5-Pro (Xiaomi), Qwen3.6 Plus (Alibaba), Nemotron 3 Super (NVIDIA).
2026-05-17Added Gemini 3.1 Pro Preview. Filled in AA-Omni scores for Claude Sonnet 4.6 (12) and GLM 5.1 (2).
2026-05-17Added AA-Omni column (AA-Omniscience Index from Artificial Analysis).
2026-05-17Added DeepSeek V4 Pro and DeepSeek V4 Flash.
2026-05-17Removed Context and Tier columns.
2026-05-17Added Kimi K2.6 (Moonshot AI) and GLM 5.1 (Z.ai). Added sort-order comment to table.
2026-05-17Initial document created. Added Claude Opus 4.7 and Claude Sonnet 4.6.