Part of Best LLM for Coding
Best LLM for Agentic Coding
Claude Fable 5.2 ranks first for Agentic Coding as of November 15, 2026.[1][2] It scores 65.76% ± 2.85 (95% interval) on Terminal-Bench 4.1 (#1 of 46) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks. Claude Opus 5.5 is the next-ranked alternative. It ranks #1 of 47 on Design Arena's web-app agent category (Elo 1321), based on blind human preference between agent-built results.
- Best open-weight pick: DeepSeek V4.1 Flash (#3)
- Compare Claude Fable 5.2 vs Claude Opus 5.5 →
About this recommendation
- Updated
- Evidence through
- Sources
- 9
- Revision
- v143
Decision audit
Why this result
Inspect the inputs and the computed order behind the recommendation.
Models screened
24
live candidates
Evaluation feeds
7
task-weighted
Winner coverage
46%
intended feed weight
Largest provider share
3 of 8
Anthropic
Anthropic: Claude Fable 5.2 leads the measured task comparison with 3 independent evaluation families. Community volume, release age and usage popularity add no ranking points.
Real-world track record is still developing. The recommendation can lead on independent measurements while practitioner reports accumulate.
Sources evaluated
These are the weights used in scoring. Missing measurements are omitted from each model’s average. A feed with no result for any model released in the last 60 days is stale and does not score until it catches up.
| Evaluation feed | Weight | Winner result | Field measured |
|---|---|---|---|
| Terminal-Bench (4.1)Observed 2026-11-15 · evaluated 2026-11-13 · Claude Code; max effort · 95% interval ±2.85 | 31% | #1 | 19/24 |
| SWE-rebenchstale | 0% | no current result | 0/24 |
| LiveBench Agentic Codingfailing; last successful check 2026-10-18 | 20% | not measured | 15/24 |
| Design Arena Full-stackObserved 2026-11-15 · evaluation date not supplied · anthropic/claude-fable-5.2-20261109 | 14% | #1 | 8/24 |
| LMArena AgentObserved 2026-11-11 · evaluation date not supplied · max effort | 13% | #1 | 22/24 |
| Design Arena Web-apps | 11% | not measured | 12/24 |
| LMArena Agent Bash Recovery | 11% | not measured | 14/24 |
Provider concentration
Each exact model is scored separately; provider identity is not a ranking input.
- Anthropic3 models
- OpenAI2 models
- deepseek1 model
- Google1 model
- Moonshot AI1 model
Decision table
Every published model is shown in computed order. Practitioner sources are distinct community threads, not the citations repeated in the prose below.
Anthropic · coverage 46%
#1 Design Arena Full-stack · #1 LMArena Agent
Anthropic · coverage 80%
#1 Design Arena Web-apps · #1 Terminal-Bench
deepseek · coverage 25%
#1 LiveBench Agentic Coding · #7 LMArena Agent Bash Recovery
Google · coverage 35%
#3 LMArena Agent · #3 Terminal-Bench
Anthropic · coverage 69%
#1 LMArena Agent Bash Recovery · #4 Design Arena Web-apps
Moonshot AI · coverage 55%
#3 Design Arena Full-stack · #3 Design Arena Web-apps
OpenAI · coverage 60%
#6 LMArena Agent · #6 Terminal-Bench
OpenAI · coverage 69%
#5 Terminal-Bench · #6 Design Arena Web-apps
Task score averages the available, task-weighted measurements. Benchmark positions compare models matched to our catalog. Models measured by at least two independent evaluator families are considered first. Missing tests, practitioner report counts, popularity and release age add no points. Coverage describes evidence breadth, not a probability of correctness. Scores within two points are shown as close under this policy; this is not a statistical confidence interval. When the leading scores are that close, the model measured by more evaluator families ranks first.
- 01
Claude Fable 5.2Anthropic · $8 in / $40 out per M tokens · 1M context
Claude Fable 5.2 scores 65.76% ± 2.85 (95% interval) on Terminal-Bench 4.1 (#1 of 46) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.
Best when: Its strongest result for this task is on Terminal-Bench, which measures completing difficult repository and terminal tasks through an agent harness.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 02
Claude Opus 5.5Anthropic · $4 in / $20 out per M tokens · 1M context
Claude Opus 5.5 ranks #1 of 47 on Design Arena's web-app agent category (Elo 1321), based on blind human preference between agent-built results.
Best when: Its strongest result for this task is on Design Arena web-app agents, which measures building web applications through an agent workflow. - 03
DeepSeek V4.1 Flashdeepseek · $0.051 in / $0.20 out per M tokens · 1.0M context · open-weight
DeepSeek V4.1 Flash scores 77.27% on LiveBench Agentic Coding (#1 of 59), covering executable JavaScript, TypeScript, and Python tasks.
Best when: Its strongest result for this task is on LiveBench Agentic Coding, which measures agentic coding tasks. - 04
Gemini 4 CarbonGoogle · $2 in / $10 out per M tokens · 1.0M context
Gemini 4 Carbon scores 64.55% ± 3.23 (95% interval) on Terminal-Bench 4.1 (#3 of 46) at high reasoning effort, using Antigravity CLI, measuring a model-and-agent configuration on reproducible terminal tasks.
Best when: Its strongest result for this task is on Terminal-Bench, which measures completing difficult repository and terminal tasks through an agent harness. - 05
Claude Sonnet 5.5Anthropic · $2 in / $10 out per M tokens · 1M context
Claude Sonnet 5.5 ranks #1 of 48 on LMArena Agent's bash-recovery signal (score 15), measuring recovery from failed shell steps.
Best when: Its strongest result for this task is on LMArena Agent bash recovery, which measures recovering from failed shell steps. - 06
Kimi K3.1 ProMoonshot AI · $0.24 in / $8.40 out per M tokens · 1.0M context · open-weight
Kimi K3.1 Pro ranks #3 of 44 on Design Arena's full-stack agent category (Elo 1321), based on blind human preference between agent-built results.
Best when: Its strongest result for this task is on Design Arena full-stack agents, which measures completing full-stack builds through an agent workflow.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 07
GPT-6 AstraOpenAI · $10 in / $50 out per M tokens · 1.1M context
GPT-6 Astra scores 60.3% ± 2.82 (95% interval) on Terminal-Bench 4.1 (#6 of 46) at max reasoning effort, using Codex, measuring a model-and-agent configuration on reproducible terminal tasks.
Best when: Its strongest result for this task is on Terminal-Bench, which measures completing difficult repository and terminal tasks through an agent harness. - 08
GPT-6.1 SolOpenAI · $2 in / $10 out per M tokens · 1.1M context
Ranks #5 on Terminal-Bench 4.1 with 60.91% and #6 on Design Arena web-app with Elo 1278, used for PO definition and architecture in development pipelines.
Best when: Select for terminal tasks where it ranks #5 of 46 with 60.91% on Terminal-Bench 4.1 at max reasoning effort.Watch out for
- Expect lower bash recovery where it ranks #20 of 48 with score 4.5 on LMArena Agent's bash-recovery signal.
Frequently asked
- What is the best model for Agentic Coding?
- Claude Fable 5.2 ranks first as of November 15, 2026. Claude Fable 5.2 scores 65.76% ± 2.85 (95% interval) on Terminal-Bench 4.1 (#1 of 46) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.[1]
- What is the best alternative to Claude Fable 5.2?
- Claude Opus 5.5 ranks second. Claude Opus 5.5 ranks #1 of 47 on Design Arena's web-app agent category (Elo 1321), based on blind human preference between agent-built results.[2]
- Which open-weight model ranks highest for Agentic Coding?
- DeepSeek V4.1 Flash is the highest-ranked open-weight model, at #3. DeepSeek V4.1 Flash scores 77.27% on LiveBench Agentic Coding (#1 of 59), covering executable JavaScript, TypeScript, and Python tasks.[3]
Sources
- 1
“Within the 46 catalog models matched to this evaluation: Scores 65.76% ± 2.85 (95% interval) on Terminal-Bench 4.1 (#1 of 46) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Nov 15, 2026 - 2
“Within the 47 catalog models matched to this evaluation: Ranks #1 of 47 on Design Arena's web-app agent category (Elo 1321), based on blind human preference between agent-built results.”
Design Arena web-app agents · Benchmark · Nov 15, 2026 - 3
“Within the 59 catalog models matched to this evaluation: Scores 77.27% on LiveBench Agentic Coding (#1 of 59), covering executable JavaScript, TypeScript, and Python tasks.”
LiveBench Agentic Coding · Benchmark · Oct 18, 2026 - 4
“Within the 46 catalog models matched to this evaluation: Scores 64.55% ± 3.23 (95% interval) on Terminal-Bench 4.1 (#3 of 46) at high reasoning effort, using Antigravity CLI, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Nov 15, 2026 - 5
“Within the 48 catalog models matched to this evaluation: Ranks #1 of 48 on LMArena Agent's bash-recovery signal (score 15), measuring recovery from failed shell steps.”
LMArena Agent bash recovery · Benchmark · Oct 8, 2026 - 6
“Within the 44 catalog models matched to this evaluation: Ranks #3 of 44 on Design Arena's full-stack agent category (Elo 1321), based on blind human preference between agent-built results.”
Design Arena full-stack agents · Benchmark · Nov 15, 2026 - 7
“Within the 46 catalog models matched to this evaluation: Scores 60.3% ± 2.82 (95% interval) on Terminal-Bench 4.1 (#6 of 46) at max reasoning effort, using Codex, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Nov 15, 2026 - 8
“Within the 46 catalog models matched to this evaluation: Scores 60.91% ± 2.97 (95% interval) on Terminal-Bench 4.1 (#5 of 46) at max reasoning effort, using Codex, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Nov 15, 2026 - 9
“Within the 48 catalog models matched to this evaluation: Ranks #20 of 48 on LMArena Agent's bash-recovery signal (score 4.5), measuring recovery from failed shell steps.”
LMArena Agent bash recovery · Benchmark · Oct 8, 2026
Rankings synthesized from community evidence and open benchmarks. See methodology. Not driven by vendor marketing.