Part of Best LLM for AI Agents
Best LLM for Hermes Agent
Claude Opus 5.5 ranks first for Hermes Agent as of October 22, 2026.[1][2][3] Leads both Terminal-Bench 4.0 and LMArena's agentic arena, scoring 64.85% on reproducible terminal tasks and ranking #1 with a 20.9 score on human-preference tool-use and multi-step performance. Claude Sonnet 5.5 is the next-ranked alternative. It scores 61.82% ± 2.94 (95% interval) on Terminal-Bench 4.0 (#2 of 23) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.
- Best open-weight pick: GLM 5.3 (#6)
- Compare Claude Opus 5.5 vs Claude Sonnet 5.5 →
About this recommendation
- Updated
- Evidence through
- Sources
- 10
- Revision
- v103
Decision audit
Why this result
Inspect the inputs and the computed order behind the recommendation.
Models screened
24
live candidates
Evaluation feeds
5
task-weighted
Winner coverage
56%
intended feed weight
Largest provider share
3 of 8
Anthropic
Anthropic: Claude Opus 5.5 leads the measured task comparison with 2 independent evaluation families. Community volume, release age and usage popularity add no ranking points. This category also weights its stated price or context criteria.
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 |
|---|---|---|---|
| Berkeley Function Callingstale | 0% | no current result | 0/24 |
| LMArena AgentObserved 2026-10-20 · evaluation date not supplied · max effort | 43% | #1 | 24/24 |
| price weight | 28% | 38/100 | 24/24 |
| Terminal-Bench (4.0)Observed 2026-10-22 · evaluated 2026-10-06 · Claude Code; max effort · 95% interval ±3.11 | 28% | #1 | 14/24 |
| SWE-rebenchstale | 0% | no current result | 0/24 |
Provider concentration
Each exact model is scored separately; provider identity is not a ranking input.
- Anthropic3 models
- OpenAI2 models
- Z.ai2 models
- xAI1 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 56%
#1 LMArena Agent · #1 Terminal-Bench
Anthropic · coverage 56%
#2 Terminal-Bench · #5 LMArena Agent
OpenAI · coverage 56%
#3 Terminal-Bench · #4 LMArena Agent
OpenAI · coverage 56%
#2 LMArena Agent · #3 Terminal-Bench
Anthropic · coverage 56%
#3 LMArena Agent · #5 Terminal-Bench
Z.ai · coverage 56%
#9 Terminal-Bench · #25 LMArena Agent
Z.ai · coverage 56%
#12 Terminal-Bench · #31 LMArena Agent
xAI · coverage 56%
#10 Terminal-Bench · #16 LMArena Agent
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 Opus 5.5Anthropic · $4 in / $20 out per M tokens · 1M context
Leads both Terminal-Bench 4.0 and LMArena's agentic arena, scoring 64.85% on reproducible terminal tasks and ranking #1 with a 20.9 score on human-preference tool-use and multi-step performance.
Best when: Use for complex terminal-based automation where highest accuracy on reproducible tasks matters, as it scores 64.85% on Terminal-Bench 4.0.More tips
- Deploy when human preference rankings for tool-use and multi-step tasks are critical, holding #1 rank with score 20.9 on LMArena's agentic arena.
No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 02
Claude Sonnet 5.5Anthropic · $2 in / $10 out per M tokens · 1M context
Claude Sonnet 5.5 scores 61.82% ± 2.94 (95% interval) on Terminal-Bench 4.0 (#2 of 23) 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.
- 03
GPT-6.1 SolOpenAI · $2 in / $10 out per M tokens · 1.1M context
Ties for #3 on Terminal-Bench 4.0 with 58.18% and ranks #4 on LMArena's agentic arena with score 13.5, showing balanced performance across terminal and multi-step tasks.
Best when: Use for multi-step tool-use workflows where human preference rankings matter, scoring 13.5 and ranking #4 on LMArena's agentic arena.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 04
GPT-6 AstraOpenAI · $10 in / $50 out per M tokens · 1.1M context
Ties for #3 on Terminal-Bench 4.0 with 58.18% and ranks #2 on LMArena's agentic arena with score 14.7, excelling at multi-step tool-use per human preference.
Best when: Prioritize for multi-step agentic workflows where human preference is key, ranking #2 with score 14.7 on LMArena's agentic arena.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 05
Claude Fable 5.1Anthropic · $10 in / $50 out per M tokens · 1M context
Ranks #5 on Terminal-Bench 4.0 with 57.88% and #3 on LMArena's agentic arena with score 14.2, offering strong multi-step performance with slightly lower terminal accuracy.
Best when: Use for balanced agentic workloads, ranking #3 with score 14.2 on LMArena's agentic arena for tool-use and multi-step tasks.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 06
GLM 5.3Z.ai · $0.079 in / $1.40 out per M tokens · 1.0M context · open-weight
Open-weight model ranking #9 on Terminal-Bench 4.0 with 41.82% and #25 on LMArena's agentic arena with score 1.4, offering self-hostable option with significant capability gaps versus closed-weight leaders.
Best when: Deploy as open-weight option for terminal tasks where 41.82% accuracy is acceptable, ranking #9 on Terminal-Bench 4.0.Watch out for
- Expect weak multi-step tool-use performance per human preference, ranking #25 with score 1.4 on LMArena's agentic arena.
- 07
GLM 5.3 FlashZ.ai · $0.006 in / $0.19 out per M tokens · 1.0M context · open-weight
GLM 5.3 Flash scores 35.76% ± 3.54 (95% interval) on Terminal-Bench 4.0 (#12 of 23) at none 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.
- 08
Grok 4.7xAI · $2 in / $6 out per M tokens · 500K context
Ranks #10 on Terminal-Bench 4.0 with 37.58% and #16 on LMArena's agentic arena with score 3.2, showing mid-tier performance on both terminal and multi-step tasks.
Best when: Use for general agentic tasks where moderate performance suffices, ranking #16 with score 3.2 on LMArena's agentic arena.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
Frequently asked
- What is the best model for Hermes Agent?
- Claude Opus 5.5 ranks first as of October 22, 2026. Leads both Terminal-Bench 4.0 and LMArena's agentic arena, scoring 64.85% on reproducible terminal tasks and ranking #1 with a 20.9 score on human-preference tool-use and multi-step performance.[1][2]
- What is the best alternative to Claude Opus 5.5?
- Claude Sonnet 5.5 ranks second. Claude Sonnet 5.5 scores 61.82% ± 2.94 (95% interval) on Terminal-Bench 4.0 (#2 of 23) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.[3]
- Which open-weight model ranks highest for Hermes Agent?
- GLM 5.3 is the highest-ranked open-weight model, at #6. Open-weight model ranking #9 on Terminal-Bench 4.0 with 41.82% and #25 on LMArena's agentic arena with score 1.4, offering self-hostable option with significant capability gaps versus closed-weight leaders.[4][5]
Sources
- 1
“Within the 23 catalog models matched to this evaluation: Scores 64.85% ± 3.11 (95% interval) on Terminal-Bench 4.0 (#1 of 23) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Oct 22, 2026 - 2
“Within the 48 catalog models matched to this evaluation: Ranks #1 of 48 on LMArena's agentic arena (score 20.9), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026 - 3
“Within the 23 catalog models matched to this evaluation: Scores 61.82% ± 2.94 (95% interval) on Terminal-Bench 4.0 (#2 of 23) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Oct 22, 2026 - 4
“Within the 23 catalog models matched to this evaluation: Scores 41.82% ± 3.23 (95% interval) on Terminal-Bench 4.0 (#9 of 23) at max reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Oct 22, 2026 - 5
“Within the 48 catalog models matched to this evaluation: Ranks #25 of 48 on LMArena's agentic arena (score 1.4), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026 - 6
“Within the 48 catalog models matched to this evaluation: Ranks #4 of 48 on LMArena's agentic arena (score 13.5), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026 - 7
“Within the 48 catalog models matched to this evaluation: Ranks #2 of 48 on LMArena's agentic arena (score 14.7), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026 - 8
“Within the 48 catalog models matched to this evaluation: Ranks #3 of 48 on LMArena's agentic arena (score 14.2), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026 - 9
“Within the 23 catalog models matched to this evaluation: Scores 35.76% ± 3.54 (95% interval) on Terminal-Bench 4.0 (#12 of 23) at none reasoning effort, using Claude Code, measuring a model-and-agent configuration on reproducible terminal tasks.”
Terminal-Bench · Benchmark · Oct 22, 2026 - 10
“Within the 48 catalog models matched to this evaluation: Ranks #16 of 48 on LMArena's agentic arena (score 3.2), measuring tool-use and multi-step task performance from human preference.”
LMArena agentic arena · Benchmark · Oct 20, 2026
Rankings synthesized from community evidence and open benchmarks. See methodology. Not driven by vendor marketing.