Use Cases Timeline Weekly Claw Benchy Subscribe Contact
Appearance
GitHub X HiM Privacy
← Back to Benchmarks
Model Detail

PrismML Bonsai 1.7B

Interesting as a tiny local experiment, but not trustworthy enough for operator work without supervision.

How it compares Source artifacts
56/100 Benchmark score
Tiny local utility model
Rank
#10 of 13
Source
PrismML local benchmark
Run date
Avg latency
8.8s
01 · Benchmark score 56/100
02 · Average latency 8.8s
03 · Role Tiny local utility model
Per-suite breakdown

Scorecard

Operator board #12

56/100 operator score

Messaging Quick execution pack

Messaging Tool Planning v2

External canon Prism ternary local model · 1.7B GGUF CPU run

Public benchmark references

Cost Local / experimental

Tiny local model, useful for light work, not a serious operator default.

Strengths
  • Very small footprint for local inference
  • Reasonable extraction on simple tasks
  • Usable concise summarization
Weaknesses
  • Invented the wrong tool path for X/Twitter threads
  • Leaked <think> tags, hurting strict JSON reliability
  • Generic and partly inverted operational reasoning
Operator read

Compact ternary local model that can do light extraction and summarization, but weak operator judgment makes it a poor default for real OpenClaw routing work.

Task breakdown

How each task scored

TaskScoreTimeWhat happened
t1 JSON extract 18/25 8s Got the extraction mostly right, but leaked chain-of-thought marker and simplified urgency, so it missed strict JSON compliance.
t2 Routing 4/25 9.3s Failed the real tool-routing test. The correct OpenClaw path is Bird first for X threads, not an invented url tool.
t3 Reasoning 9/25 11.1s Partly recognized on-prem/privacy reasons, but inverted the defaulting logic and stayed generic instead of concrete agent-routing guidance.
t4 Summary 25/25 6.8s Clean and accurate summary within the length budget.
Best output25/25
t4 Summary
Gemma 4 31B has strong raw benchmarks, but Gemma 4 26B MoE is better for local deployment due to its headroom for concurrency, latency, and cost, while keeping useful quality.
Worst output4/25
t2 Routing
Use openclaw’s url tool to fetch the thread data from the URL, which is the best tool path for reading a single X/Twitter thread. The url tool handles parsing and fetching the thread content, making it the most efficient and reliable method for reading and summarizing the benchmarks.
Bottom line

Bonsai 1.7B is usable for lightweight local extraction and summarization, but it is not trustworthy enough for routing, benchmark-sensitive judgment, or default operator work.

MFSF
MFSF
  • This host was only 4 vCPU, 15 GiB RAM, and CPU-only, so Bonsai 8B was not practical to benchmark interactively here.
  • Prism CLI defaults to interactive chat formatting and exposes thinking tags, which hurts strict benchmark compliance unless post-processed.
Comparison

Where PrismML Bonsai 1.7B lands

OperatorIndex
  1. 01 Gemini 3.1 ProMessaging benchmark + external canon 100
  2. 01 Claude Sonnet 4.6Messaging benchmark + external canon 100
  3. 01 Gemini FlashMessaging benchmark canon 100
  4. 04 Claude Opus 4.6Operator Suite v2 95.3
  5. 05 GLM-5-TurboOperator Suite v2 95
  6. 06 Gemma 4 31BEnterprise Ollama 92
  7. 07 MiniMax M2.7Operator Suite v2 90.8
  8. 08 GPT-5.4External benchmark canon 88
  9. 09 Gemma 4 26BEnterprise Ollama 80
  10. 10 PrismML Bonsai 1.7BPrismML local benchmark 56
  11. 11 Qwen 3.6 27B NVFP4Enterprise oMLX 40
  12. 11 Qwen 3.6 27B MXFP4Enterprise oMLX 40
  13. 13 Qwen 3.6 27B 4bitEnterprise oMLX 20
Source artifacts

Raw machine-readable files for anyone who wants to dig deeper or run their own analysis.