Benchmarks

Compare AI models

Pick two models to see release dates, benchmark scores, and community ratings side by side.

Frontier benchmarks

Frontier board

6 releases. Five ways to compare.

Selected for recent releases and benchmark coverage, not a single overall rank. Each column sorts the same models by its own score. No reading means no bar.

○ Independent evaluation · ◆ Vendor or self-reported, labeled per reading.

Coding

BenchLM · 0 to 100

  1. Claude Opus 5.582.71Independent

    Anthropic

  2. Claude Fable 5.176.84Independent

    Anthropic

  3. GPT-6 Astra75.54Independent

    OpenAI

  4. Gemini 4 Argon73.04Independent

    Google DeepMind

  5. GPT-6.1 Sol71.02Independent

    OpenAI

  6. Muse Spark Open56.11Independent

    Meta

Agentic

BenchLM · 0 to 100

  1. Claude Opus 5.590.37Independent

    Anthropic

  2. Claude Fable 5.180.77Independent

    Anthropic

  3. Gemini 4 Argon78.7Independent

    Google DeepMind

  4. GPT-6.1 Sol76.19Independent

    OpenAI

  5. GPT-6 Astra75.07Independent

    OpenAI

  6. Muse Spark Open60.14Independent

    Meta

Reasoning

BenchLM · 0 to 100

  1. GPT-6 Astra93.3Independent

    OpenAI

  2. GPT-6.1 Sol90.5Independent

    OpenAI

  3. Gemini 4 Argon86.9Independent

    Google DeepMind

  4. Claude Opus 5.586.1Independent

    Anthropic

  5. Claude Fable 5.185.2Independent

    Anthropic

  6. Muse Spark OpenNot reported

    Meta

Knowledge

BenchLM · 0 to 100

  1. Claude Opus 5.590.27Independent

    Anthropic

  2. Claude Fable 5.186.89Independent

    Anthropic

  3. GPT-6 Astra86.26Independent

    OpenAI

  4. GPT-6.1 Sol86.1Independent

    OpenAI

  5. Gemini 4 Argon81.41Independent

    Google DeepMind

  6. Muse Spark Open64.71Independent

    Meta

Arena

LM Arena · Elo

  1. Google DeepMind

  2. Anthropic

  3. Anthropic

  4. GPT-6.1 Sol1,479.36Independent

    OpenAI

  5. GPT-6 Astra1,475.5Independent

    OpenAI

  6. Muse Spark OpenNot reported

    Meta

Score links open the source; hover for capture details. Bar lengths start at zero. Arena lengths use the full public board maximum.

The shape of the scores

First three featured models. Toggle a name to isolate a line. Missing readings leave gaps, not zeroes.

CodingAgenticReasoningKnowledgeArena050100Claude Opus 5.5, Coding: 82.71. Independent. Captured 2026-10-21.Claude Opus 5.5, Agentic: 90.37. Independent. Captured 2026-10-21.Claude Opus 5.5, Reasoning: 86.1. Independent. Captured 2026-10-21.Claude Opus 5.5, Knowledge: 90.27. Independent. Captured 2026-10-21.Claude Opus 5.5, Arena: 1,508.75. Independent. Captured 2026-10-21.Claude Fable 5.1, Coding: 76.84. Independent. Captured 2026-10-21.Claude Fable 5.1, Agentic: 80.77. Independent. Captured 2026-10-21.Claude Fable 5.1, Reasoning: 85.2. Independent. Captured 2026-10-21.Claude Fable 5.1, Knowledge: 86.89. Independent. Captured 2026-10-21.Claude Fable 5.1, Arena: 1,502.58. Independent. Captured 2026-10-21.GPT-6 Astra, Coding: 75.54. Independent. Captured 2026-10-21.GPT-6 Astra, Agentic: 75.07. Independent. Captured 2026-10-21.GPT-6 Astra, Reasoning: 93.3. Independent. Captured 2026-10-21.GPT-6 Astra, Knowledge: 86.26. Independent. Captured 2026-10-21.GPT-6 Astra, Arena: 1,475.5. Independent. Captured 2026-10-21.

3 of 3 series shown.

Axes share drawing space, not meaning. Arena is Elo divided by the current board maximum; it is not a percentage accuracy.

The newest arrival

Muse Spark Open

Released 2026-10-15 · At least two readings

No predecessor is linked in the catalog yet. We do not infer a generation jump from a name.

Latest release with linked lineage

Claude Fable 5.1 vs Claude Fable 5

Coding
+5.54pts
71.3 to 76.84 Independent2026-10-21 / 2026-10-21
Agentic
+4.95pts
75.82 to 80.77 Independent2026-10-21 / 2026-10-21
Knowledge
+4.29pts
82.6 to 86.89 Independent2026-10-21 / 2026-10-21
Arena
-1.08Elo
1,503.66 to 1,502.58 Independent2026-10-21 / 2026-10-21

Latest snapshots, not a controlled before-and-after trial. Capture dates and evaluation setups may differ.

Scores and sources

Frontier board: original scores. Not reported is missing, not zero.
ModelMetricScore / provenanceCapturedSource label / votes
Claude Opus 5.5CodingBenchLM.ai82.71Independent2026-10-21Claude Opus 5.5
Claude Fable 5.1CodingBenchLM.ai76.84Independent2026-10-21Claude Fable 5.1
GPT-6 AstraCodingBenchLM.ai75.54Independent2026-10-21GPT-6 Astra
Gemini 4 ArgonCodingBenchLM.ai73.04Independent2026-10-21Gemini 4 Argon
GPT-6.1 SolCodingBenchLM.ai71.02Independent2026-10-21GPT-6.1 Sol
Muse Spark OpenCodingBenchLM.ai56.11Independent2026-10-21Muse Spark Open
Claude Opus 5.5AgenticBenchLM.ai90.37Independent2026-10-21Claude Opus 5.5
Claude Fable 5.1AgenticBenchLM.ai80.77Independent2026-10-21Claude Fable 5.1
Gemini 4 ArgonAgenticBenchLM.ai78.7Independent2026-10-21Gemini 4 Argon
GPT-6.1 SolAgenticBenchLM.ai76.19Independent2026-10-21GPT-6.1 Sol
GPT-6 AstraAgenticBenchLM.ai75.07Independent2026-10-21GPT-6 Astra
Muse Spark OpenAgenticBenchLM.ai60.14Independent2026-10-21Muse Spark Open
GPT-6 AstraReasoningBenchLM.ai93.3Independent2026-10-21GPT-6 Astra
GPT-6.1 SolReasoningBenchLM.ai90.5Independent2026-10-21GPT-6.1 Sol
Gemini 4 ArgonReasoningBenchLM.ai86.9Independent2026-10-21Gemini 4 Argon
Claude Opus 5.5ReasoningBenchLM.ai86.1Independent2026-10-21Claude Opus 5.5
Claude Fable 5.1ReasoningBenchLM.ai85.2Independent2026-10-21Claude Fable 5.1
Muse Spark OpenReasoningBenchLM.aiNot reportedNot reportedNot reported
Claude Opus 5.5KnowledgeBenchLM.ai90.27Independent2026-10-21Claude Opus 5.5
Claude Fable 5.1KnowledgeBenchLM.ai86.89Independent2026-10-21Claude Fable 5.1
GPT-6 AstraKnowledgeBenchLM.ai86.26Independent2026-10-21GPT-6 Astra
GPT-6.1 SolKnowledgeBenchLM.ai86.1Independent2026-10-21GPT-6.1 Sol
Gemini 4 ArgonKnowledgeBenchLM.ai81.41Independent2026-10-21Gemini 4 Argon
Muse Spark OpenKnowledgeBenchLM.ai64.71Independent2026-10-21Muse Spark Open
Gemini 4 ArgonArenaLM Arena1,523.93Independent2026-10-21Gemini 4 Argon (High)4,900 votes
Claude Opus 5.5ArenaLM Arena1,508.75Independent2026-10-21Claude Opus 5.5 (High)9,496 votes
Claude Fable 5.1ArenaLM Arena1,502.58Independent2026-10-21Claude Fable 5.1 (Max)15,623 votes
GPT-6.1 SolArenaLM Arena1,479.36Independent2026-10-21GPT-6.1 Sol (Max)7,453 votes
GPT-6 AstraArenaLM Arena1,475.5Independent2026-10-21GPT-6 Astra (Max)13,293 votes
Muse Spark OpenArenaLM ArenaNot reportedNot reportedNot reported
Additional readings for these models. Separate benchmarks, not substitutes for the five chart axes.
ModelMetric / sourceScore / provenanceCaptured
Claude Opus 5.5MultimodalGroundedBenchLM.ai92.2Independent2026-10-21
Claude Opus 5.5GDP.pdfEpoch AI30.6Independent2026-10-21
Claude Fable 5.1Humanity's Last ExamEpoch AI46.5Independent2026-10-21
Claude Fable 5.1GDP.pdfEpoch AI29.6Independent2026-10-21
GPT-6 AstraHumanity's Last ExamEpoch AI54.8Independent2026-10-21
Open page →
Provider
Mistral AI
-
Out yet?
No
-
Status
Speculated
-
Release date
-
-
Vibe rating?OutYet's community rating: signed-in users score the model 1–10. Shown as the average and the number of votes.
-
-
Predecessor
-
-
Successor
-
-