RankingTerminal-Bench 2.1 · 2.1

Terminal-Bench 2.1 · 2.1

Data updated 12 Sept 2026

Bucket
Coding
Unit
percent
Direction
Higher is better
Version
2.1
Display harness
moonshotai/Kimi-K3
Board
https://huggingface.co/moonshotai/Kimi-K3

Compare published benchmark results with category weights →

Models

ModelScoreHarnessEvidenceSource-recorded date
GPT-5.6 SolOpenAI88.8%
Reported settings & source

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

moonshotai/Kimi-K3lab self-report2026-09-12
Kimi K3Moonshot88.3%
Reported settings & source

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: Terminal-Bench 2.1 · reviewed 2026-09-12

moonshotai/Kimi-K3lab self-report2026-09-12
Claude Fable 5Anthropic88%
Reported settings & source

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

moonshotai/Kimi-K3lab self-report2026-09-12
Claude Opus 4.8Anthropic84.6%
Reported settings & source

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

moonshotai/Kimi-K3lab self-report2026-09-12
GPT-5.5OpenAI83.4%
Reported settings & source

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

moonshotai/Kimi-K3lab self-report2026-09-12
GLM-5.2Z.ai82.7%
Reported settings & source

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

moonshotai/Kimi-K3lab self-report2026-09-12
Claude Opus 4.7Anthropic66.1%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
Claude Opus 4.8Anthropic82.7%
Reported settings & source

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

Muse Spark 1.1 evaluation report Figure44lab self-report2026-09-06
Gemini 3.1 ProGoogle70.3%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
Gemini 3.1 ProGoogle70.3%
Reported settings & source

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

Muse Spark 1.1 evaluation report Figure44lab self-report2026-09-06
GLM-5.1Z.ai48.3%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
GPT-5.5OpenAI78.2%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
GPT-5.5OpenAI83.4%
Reported settings & source

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

Muse Spark 1.1 evaluation report Figure44lab self-report2026-09-06
Kimi-K2.6Moonshot53.9%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
MiniMax-M2.7MiniMax51.1%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
MiniMax-M3MiniMax66%
Reported settings & source

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 M3 model card benchmark figurelab self-report2026-09-06
Muse Spark 1.1Meta80%
Reported settings & source

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

Muse Spark 1.1 evaluation report Figure44lab self-report2026-09-06