Skip to content
AI Atlas
BenchmarkActivecategory · agenticfamily · terminal-bench · variant 1.0

Terminal-Bench

tbench.ai

terminal tasks solved by agents

quality57

Updated 12 min ago · first seen 11 Sept 2026

Metric
accuracy · % ↑
Current results
821
Models
374
Current leader
gpt-5.6-sol 65.9%

Score history · Reka Flash 3 1 row

Not enough history to chart — a single observation (0% on 11 Sept 2026). Rows under different configurations count separately; the list below shows each one.

  • 0%aa_slug=reka-flash-3 · variant=hard · evaluator=Artificial Analysis · index_version=4.311 Sept 2026

Back to the leaderboard

Frontier over time · accuracy · variant=hard · evaluator=Artificial Analysis

10 leader changes recorded, all dated 11 Sept 2026 — the frontier line needs at least two distinct dates. The corpus is young: every result was first observed on the same day, so leader changes will separate in time as sources are re-crawled.

  1. 65.9%gpt-5.6-sol OpenAI Independent11 Sept 2026
  2. 61.4%gpt-5.6-sol OpenAI Independent11 Sept 2026
  3. 53.0%Claude Sonnet 4.6 Anthropic Independent11 Sept 2026
  4. 51.5%Claude Opus 4.7 Anthropic Independent11 Sept 2026
  5. 50.8%Z.ai GLM 5.2 Z.ai (Zhipu AI) Independent11 Sept 2026
  6. 49.2%KAT-Coder-Pro V2 Kwaipilot Independent11 Sept 2026
  7. 33.3%GPT-5.1-Codex Mini OpenAI Independent11 Sept 2026
  8. 26.5%Grok 4.3 xAI Independent11 Sept 2026

Includes closed rows (history). A point is emitted whenever a result beats every earlier result of the same group, ordered by evaluated_at when the source publishes it, else observed_at.

Leaderboard 315 models

Select models with +, then Compare.

Leaderboard
#ModelScoreTrustConfigurationvs leaderEvaluatedSourceActions
276lfm2-8b-a1bOpen weightsLiquid AI · LFM20%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276llama-3-3-nemotron-super-49bOpen weightsNVIDIA · Llama 3.3 · best of 2 rows0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276llama-3-instruct-8bOpen weightsMeta AI · Llama 30%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276minicpm-v4-6-1-3bOpen weightsOpenBMB0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276minicpm5-1bOpen weightsOpenBMB · best of 2 rows0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276molmo-7b-dOpen weightsAllen Institute for AI · Molmo0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276nanbeige4-1-3bOpen weightsNanbeige0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276olmo-2-32bOpen weightsAllen Institute for AI · OLMo 20%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276olmo-2-7bOpen weightsAllen Institute for AI · OLMo 20%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276olmo-3-1-32b-instructOpen weightsAllen Institute for AI · OLMo 3.1 · best of 2 rows0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276phi-3-miniOpen weightsMicrosoft · Phi30%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276phi-4-miniOpen weightsMicrosoft · Phi40%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276qwen3-0.6b-instructOpen weightsAlibaba Group · Qwen3.00%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276qwen3-1.7b-instructOpen weightsAlibaba Group · Qwen3.1 · best of 2 rows0%IndependentreasoningonPartially comparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History
276tiny-aya-globalRestricted weightsCohere · Aya0%Independentgroup defaultsComparable0.00 ptobs. 11 Sept 2026artificialanalysis.aiT2History

One row per canonical model — its best current row inside this comparability group (effort variants are folded into the model). Bars are relative to the page's best score. “vs leader” reads comparability: partially comparable = same task, conditions differ (reasoning effort, temperature, judge). Rules →