GPT-5.5

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: AA-Briefcase (Elo) · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: AA-LCR · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / Agents' Last Exam / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites official Agents Last Exam leaderboard; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: Agents' Last Exam · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis / APEX-Agents leaderboard; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: APEX-Agents · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Abstract Reasoning table / ARC-AGI-3 / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Coding table / Artificial Analysis Coding Agent Index v1.1 / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / Artificial Analysis Intelligence Index v4.1 / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Tool Use table / AutomationBench / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: AutomationBench · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: BabyVision w/ python · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Computer Use table / BenchCAD / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; Python tool enabled

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Computer Use table / BenchCAD (python tool) / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / Big Finance Bench / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Computer Use table / BrowseComp / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Anthropic or OpenAI release reports; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: BrowseComp · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; reduced or absent production safeguards

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Cybersecurity table / Capture-the-Flag Challenges / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: CharXiv (RQ) · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: CharXiv (RQ) · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Vals AI; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: CorpFin v2 · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: CritPt · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: DeepSWE · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Coding table / DeepSWE v1.1 / GPT‑5.5 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; reduced or absent production safeguards; ExploitBench API harness; five seeds; reasoning continuity

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Cybersecurity table / ExploitBench / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; reduced or absent production safeguards; six-hour evaluation cap; alpha API latency rescaled to public API

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Cybersecurity table / ExploitGym / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Vals AI; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: Finance Agent v2 · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Academic table / FrontierMath Tier 1-3 (v2) / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Academic table / FrontierMath Tier 4 (v2) / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites FrontierSWE leaderboard; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: FrontierSWE · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Multimodal table / gdp.pdf / GPT‑5.5 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / GDPval-AA v2 / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: GDPval-AA v2 (Elo) · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / GeneBench Pro / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Academic table / GPQA Diamond / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: GPQA Diamond · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified; BFS 1mil f1

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / GraphWalks BFS 1mil f1 / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; BFS 256k f1

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / GraphWalks BFS 256k f1 / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: Harvey Lab-AA · reviewed 2026-09-12

length-adjusted; official HealthBench scoring

Mean response length 2313 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench / gpt-5.5 / length-adjusted · reviewed 2026-09-06

unadjusted; official HealthBench scoring

Mean response length 2313 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench / gpt-5.5 / unadjusted · reviewed 2026-09-06

length-adjusted; official HealthBench scoring

Mean response length 2259 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Consensus / gpt-5.5 / length-adjusted · reviewed 2026-09-06

unadjusted; official HealthBench scoring

Mean response length 2259 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Consensus / gpt-5.5 / unadjusted · reviewed 2026-09-06

length-adjusted; official HealthBench scoring

Mean response length 2289 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Hard / gpt-5.5 / length-adjusted · reviewed 2026-09-06

unadjusted; official HealthBench scoring

Mean response length 2289 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Hard / gpt-5.5 / unadjusted · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; official paper scoring; length-adjusted

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / HealthBench Professional / GPT‑5.5 · reviewed 2026-09-06

length-adjusted; official HealthBench scoring

Mean response length 3818 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Professional / gpt-5.5 / length-adjusted · reviewed 2026-09-06

unadjusted; official HealthBench scoring

Mean response length 3818 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations.

GPT-6 Astra System Card · Table6 / HealthBench Professional / gpt-5.5 / unadjusted · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: HLE-Full · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: HLE-Full · reviewed 2026-09-12

MiniMax points out of42; comparator percentages as printed

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / Internal Research Debugging Evaluation / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: JobBench · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / KernelGen 1P / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: Kimi Code Bench 2.0 · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Vals AI; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: Legal Research Bench · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / LifeSciBench / GPT‑5.5 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / Management Consulting Tasks (Internal) / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MathVision · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MathVision · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MCP-Atlas · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MCPMark-Verified · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / MedChemBench (Internal) / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MLS-Bench-Lite · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified; no tools

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Multimodal table / MMMU Pro (no tools) / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; tools enabled

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Multimodal table / MMMU Pro (with tools) / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MMMU-Pro · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MMMU-Pro · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: MMVU · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / NanoGPT / GPT‑5.5 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: OfficeQA Pro · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: OmniDocBench · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; 8-needle 256K-512K

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / OpenAI MRCR v2 8-needle 256K-512K / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified; 8-needle 512K-1M

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / OpenAI MRCR v2 8-needle 512K-1M / GPT‑5.5 · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Computer Use table / OSWorld 2.0 / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: OSWorld 2.0 · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: OSWorld-Verified · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: PerceptionBench · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites official PostTrainBench results; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: PostTrainBench · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / PostTrainBench Lite / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites GLM release blog or Vals AI, per model; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: ProgramBench · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: ResearchRubrics · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / RSI Index / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: SaaS-Bench · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: SciCode · reviewed 2026-09-12

Launch table reported configuration; per-cell reasoning effort unspecified; reduced or absent production safeguards; public grader; May2026 JavaScript subset

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Cybersecurity table / SEC-Bench Pro / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: SpreadsheetBench 2 · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Coding table / SWE-Bench Pro / GPT‑5.5 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: SWE-Marathon · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Coding table / Terminal-Bench 2.1 / GPT‑5.5 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites GLM release blog, Artificial Analysis or OpenAI, per model; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: Terminal-Bench 2.1 · reviewed 2026-09-12

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Launch table reported configuration; per-cell reasoning effort unspecified

Provider-published result; comparator measurements are not automatically independently reproduced.

GPT-5.6: Frontier intelligence that scales with your ambition · Tool Use table / Toolathlon / GPT‑5.5 · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: Toolathlon-Verified · reviewed 2026-09-12

MiniMax points out of42; comparator percentages as printed

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: Video-MME (w. sub) · reviewed 2026-09-12

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

MiniMax launch figure methodology. SWE internal ClaudeCode 4 runs; Terminal 8CPU16GB2h128k Terminus2; Paper/PostTrain Ralph-loop12h; GDPval internal rubric grader; BrowseComp discard64k; OSWorld361tasks. Comparator provider/leaderboard scores where footnoted.

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Meta Model API xhigh; Gemini high; Opus max; GPT xhigh. Best of self-reported or Meta reproduction unless specified. OSWorld GUI only; Terminal bash only5attempts89tasks6CPU8GB; DeepSWE no internet5attempts.

First-party reported result; comparator results retain the source evaluation setup.

Muse Spark 1.1 evaluation report Figure44 · Figure44 physical page101; methodology pages101–105 · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: WorldVQA ForceAnswer · reviewed 2026-09-12

Launch figure; agent final assets

First-party reported result; comparator results retain the source evaluation setup.

MiniMax M3 model card benchmark figure · figures/benchmark.jpeg · reviewed 2026-09-06

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. without tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: ZeroBench (pass@5) · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. with tools Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

First-party reported result.

moonshotai/Kimi-K3 · Performance table: ZeroBench (pass@5) · reviewed 2026-09-12

Table headers: Kimi K3, GPT-5.6 Sol, Opus 4.8 and GLM-5.2 max; GPT-5.5 xhigh; Fable 5 max with fallbacks. Benchmark-specific footnotes override header settings; cited results are not new Moonshot evaluations. Benchmark-specific tools, harness and budget documented under Evaluation Details. Kimi temp1; single-step top_p.95,agentic top_p1; vision tools=Python,3runs except ZeroBench5runs.

Cited result; see benchmark-specific Evaluation Details.

Comparison limit: Kimi K3 Evaluation Details cites Artificial Analysis; retain as published context, not a new Moonshot comparison.

moonshotai/Kimi-K3 · Performance table: τ³-Banking · reviewed 2026-09-12