feat: add filesystem benchmark mode

[HeyGarrison]

Summary

• add a new fs benchmark mode that runs in-sandbox filesystem workloads (large-file read/write, small-file churn, and metadata operations) per provider
• wire FS benchmarking into the CLI flow, result persistence (results/fs + latest.json), and composite scoring/success-rate reporting
• document the new mode in README and methodology and add npm scripts for common FS benchmark invocations

Validation

• attempted local smoke test with npm run bench:fs -- --iterations 1 --provider e2b
• environment in this workspace blocked execution (tsx: Permission denied and missing local deps), so runtime validation is expected in CI where keys and deps are configured

[github-actions]

Browser Benchmark Results

#	Provider	Score	Create	Connect	Navigate	Release	Total	Status
1	Browserbase	94.5	0.22s	0.12s	0.06s	0.13s	0.55s	10/10
2	Kernel	93.5	0.07s	0.40s	0.15s	0.07s	0.71s	10/10
3	Hyperbrowser	91.1	0.14s	0.19s	0.11s	0.16s	0.63s	10/10

---

View full run · SVG available as build artifact

[github-actions]

Sandbox Benchmark Results

Sequential

#	Provider	Score	Median TTI	P95	P99	Status
1	declaw	99.3	0.04s	0.11s	0.11s	10/10
2	daytona	98.2	0.12s	0.26s	0.26s	10/10
3	upstash	95.2	0.40s	0.59s	0.59s	10/10
4	e2b	94.4	0.46s	0.70s	0.70s	10/10
5	vercel	93.5	0.56s	0.79s	0.79s	10/10
6	blaxel	92.9	0.51s	1.02s	1.02s	10/10
7	hopx	83.1	1.32s	2.25s	2.25s	10/10
8	modal	82.2	1.59s	2.06s	2.06s	10/10
9	namespace	81.2	1.83s	1.96s	1.96s	10/10
10	cloudflare	78.3	2.01s	2.40s	2.40s	10/10
11	runloop	75.5	1.80s	3.41s	3.41s	10/10
12	codesandbox	71.2	2.72s	3.13s	3.13s	10/10
13	archil	9.4	0.57s	0.57s	0.57s	1/10

Staggered

#	Provider	Score	Median TTI	P95	P99	Status
1	daytona	98.4	0.14s	0.20s	0.20s	10/10
2	declaw	98.0	0.04s	0.43s	0.43s	10/10
3	upstash	96.1	0.37s	0.41s	0.41s	10/10
4	blaxel	94.7	0.51s	0.56s	0.56s	10/10
5	e2b	93.8	0.48s	0.83s	0.83s	10/10
6	vercel	92.3	0.69s	0.88s	0.88s	10/10
7	hopx	84.9	1.37s	1.73s	1.73s	10/10
8	namespace	79.8	1.94s	2.14s	2.14s	10/10
9	modal	79.6	1.76s	2.45s	2.45s	10/10
10	cloudflare	75.9	2.11s	2.88s	2.88s	10/10
11	runloop	74.6	1.88s	3.52s	3.52s	10/10
12	codesandbox	71.7	2.73s	2.97s	2.97s	10/10
13	archil	9.7	0.28s	0.28s	0.28s	1/10

Burst

#	Provider	Score	Median TTI	P95	P99	Status
1	declaw	98.6	0.05s	0.28s	0.28s	10/10
2	daytona	98.6	0.13s	0.15s	0.15s	10/10
3	upstash	95.2	0.45s	0.53s	0.53s	10/10
4	e2b	94.7	0.50s	0.59s	0.59s	10/10
5	blaxel	94.3	0.56s	0.58s	0.58s	10/10
6	vercel	93.1	0.58s	0.85s	0.85s	10/10
7	modal	79.5	1.87s	2.32s	2.32s	10/10
8	namespace	79.3	1.91s	2.30s	2.30s	10/10
9	hopx	78.4	1.96s	2.46s	2.46s	10/10
10	cloudflare	75.8	2.06s	2.97s	2.97s	10/10
11	runloop	70.9	2.02s	4.25s	4.25s	10/10
12	codesandbox	70.8	2.82s	3.07s	3.07s	10/10
13	archil	9.7	0.29s	0.29s	0.29s	1/10

---

View full run · SVGs available as build artifacts

[github-actions]

Storage Benchmark Results

1MB Files

#	Provider	Score	Download	Throughput	Upload	Status
1	Cloudflare R2	94.8	0.11s	78.5 Mbps	0.18s	1000/1000
2	AWS S3	94.2	0.34s	24.4 Mbps	0.06s	1000/1000
3	Tigris	93.5	0.52s	16.2 Mbps	0.19s	1000/1000

4MB Files

#	Provider	Score	Download	Throughput	Upload	Status
1	Cloudflare R2	94.9	0.18s	186.5 Mbps	0.32s	1000/1000

10MB Files

#	Provider	Score	Download	Throughput	Upload	Status
1	Cloudflare R2	94.1	0.34s	248.2 Mbps	0.81s	100/100
2	AWS S3	93.2	0.60s	139.9 Mbps	0.41s	100/100
3	Tigris	90.0	1.24s	67.7 Mbps	1.18s	100/100

16MB Files

#	Provider	Score	Download	Throughput	Upload	Status
1	Cloudflare R2	93.8	0.49s	276.4 Mbps	0.82s	100/100
2	AWS S3	90.5	1.53s	87.9 Mbps	0.43s	100/100
3	Tigris	88.4	1.72s	78.2 Mbps	1.22s	100/100

---

View full run · SVGs available as build artifacts

[github-actions]

Filesystem Benchmark Results

#	Provider	Score	Read	Write	Small Files	Metadata	Status
1	namespace	99.8	0.04s	0.04s	0.08s	0.00s	5/5
2	upstash	99.8	0.02s	0.01s	0.19s	0.01s	5/5
3	archil	99.7	0.04s	0.06s	0.21s	0.01s	5/5
4	vercel	99.7	0.07s	0.06s	0.17s	0.01s	5/5
5	blaxel	99.7	0.08s	0.06s	0.17s	0.01s	5/5
6	codesandbox	99.7	0.07s	0.06s	0.19s	0.00s	5/5
7	runloop	99.7	0.06s	0.05s	0.24s	0.01s	5/5
8	daytona	99.7	0.04s	0.01s	0.34s	0.01s	5/5
9	e2b	99.6	0.06s	0.07s	0.37s	0.01s	5/5
10	declaw	99.5	0.10s	0.11s	0.30s	0.01s	5/5
11	modal	99.3	0.06s	0.07s	0.66s	0.02s	5/5
12	cloudflare	97.9	0.45s	0.57s	1.08s	0.04s	5/5
13	hopx	93.6	2.19s	1.73s	1.72s	0.03s	5/5

---

View full run