Benchmark Almanac

Firecrawl Search Bench

Claude Opus 5.5 79% leads since November 2026
Not in model cards

Agents answer hard research questions with web search. A deterministic grader checks each answer against reference facts.

Firecrawl reports 429 questions and scores for 10 models. The page reports scores from 88.5% to 96.2% for ranked models and publishes a separate FSB-only table.

Made by
Firecrawl
Released
November 2026
Kind
Deep research
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Agentic
Environment
Browser
Refresh
One fixed set
Run by
A host
In views
Agents
Strongest model tested
Claude Opus 5.5 #5 on the capability index
Last model added
October 2026
Scores refreshed
November 21st, 2026
10
Entries on its own leaderboard, 9 models we track

Claude Opus 5.5 leads with 79% Claude Opus 5.5 leads with 79%

Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

  1. 1 Claude Opus 5.5 79%
  2. 2 GPT-6.1 Sol 77.9%
  3. 3 Kimi K3.1 Pro 77.4%
  4. 4 Pareto 26.9 77.1%
  5. 5 DeepSeek V4.1 Flash 1017 68.9%
  6. 6 Mercury 3 68.7%
  7. 7 MiniMax-M3.1-Flash 64.8%
  8. 8 GLM-5.3-Flash 61.9%
  9. 9 Claude Haiku 5.5 61.9%
  10. 10 Nemotron 3.5 Lightning 30B-A3B 44%

Source: Official Firecrawl Search Bench leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5.5 79% Sep 2026 Official leaderboard
2 GPT-6.1 Sol 77.9% Sep 2026 Official leaderboard
3 Kimi K3.1 Pro 77.4% Oct 2026 Official leaderboard
4 Pareto 26.9 77.1% Official leaderboard
5 DeepSeek V4.1 Flash 1017 68.9% Oct 2026 Official leaderboard
6 Mercury 3 68.7% Oct 2026 Official leaderboard
7 MiniMax-M3.1-Flash 64.8% Oct 2026 Official leaderboard
8 GLM-5.3-Flash 61.9% Aug 2026 Official leaderboard
9 Claude Haiku 5.5 61.9% Oct 2026 Official leaderboard
10 Nemotron 3.5 Lightning 30B-A3B 44% Aug 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 5.5 79% Sep 2026 Official leaderboard
2 GPT-6.1 Sol 77.9% Sep 2026 Official leaderboard
3 Kimi K3.1 Pro 77.4% Oct 2026 Official leaderboard
4 DeepSeek V4.1 Flash 1017 68.9% Oct 2026 Official leaderboard
5 Mercury 3 68.7% Oct 2026 Official leaderboard
6 MiniMax-M3.1-Flash 64.8% Oct 2026 Official leaderboard
7 GLM-5.3-Flash 61.9% Aug 2026 Official leaderboard
8 Claude Haiku 5.5 61.9% Oct 2026 Official leaderboard
9 Nemotron 3.5 Lightning 30B-A3B 44% Aug 2026 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: Search lift and Without Firecrawl Alexandria. Each ranks models on one part of the test. GPT-6.1 Sol leads Search lift and Kimi K3.1 Pro leads Without Firecrawl Alexandria. Arrows show how far a model moved from its overall rank among the models we track.

Search lift

9 models we track have a score here.

  1. 1 GPT-6.1 Sol ↑1 45.6%
  2. 2 Claude Haiku 5.5 ↑6 44.4%
  3. 3 Mercury 3 ↑2 42.4%
  4. 4 Claude Opus 5.5 ↓3 41.9%
  5. 5 MiniMax-M3.1-Flash ↑1 40.5%
  6. 6 Kimi K3.1 Pro ↓3 38.9%
  7. 7 DeepSeek V4.1 Flash 1017 ↓3 38%
  8. 8 GLM-5.3-Flash ↓1 36.7%
  9. 9 Nemotron 3.5 Lightning 30B-A3B 32.8%

The best score, 79%, costs $0.37 a task

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

NVIDIA Anthropic OpenAI MiniMax Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.01 $0.02 $0.05 $0.10 $0.20 $0.50 $1.00 Cost per task, log scale Claude Opus 5.5 GPT-6.1 Sol MiniMax-M3.1-Flash Nemotron 3.5 Lightning 30B-A3B 0% 20% 40% 60% 80% 100% $0.01 $0.10 $1.00 Cost per task, log scale Claude Opus 5.5 MiniMax-M3.1-Flash

Source: Official Firecrawl Search Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5.5 79% $0.37
GPT-6.1 Sol 77.9% $0.21
Kimi K3.1 Pro 77.4% $0.53
DeepSeek V4.1 Flash 1017 68.9% $0.29
Mercury 3 68.7% $0.24
MiniMax-M3.1-Flash 64.8% $0.20
GLM-5.3-Flash 61.9% $0.17
Claude Haiku 5.5 61.9% $0.17
Nemotron 3.5 Lightning 30B-A3B 44% $0.09

The best score, 79%, costs $0.37 a task

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

NVIDIA Anthropic OpenAI MiniMax Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.01 $0.02 $0.05 $0.10 $0.20 $0.50 $1.00 Cost per task, log scale Claude Opus 5.5 GPT-6.1 Sol MiniMax-M3.1-Flash Nemotron 3.5 Lightning 30B-A3B 0% 20% 40% 60% 80% 100% $0.01 $0.10 $1.00 Cost per task, log scale Claude Opus 5.5 MiniMax-M3.1-Flash

Source: Official Firecrawl Search Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5.5 79% $0.37
GPT-6.1 Sol 77.9% $0.21
Kimi K3.1 Pro 77.4% $0.53
DeepSeek V4.1 Flash 1017 68.9% $0.29
Mercury 3 68.7% $0.24
MiniMax-M3.1-Flash 64.8% $0.20
GLM-5.3-Flash 61.9% $0.17
Claude Haiku 5.5 61.9% $0.17
Nemotron 3.5 Lightning 30B-A3B 44% $0.09

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic
Own leaderboard 1 leader

Source: Official Firecrawl Search Bench leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Opus 5.5 GLM-5.3-Flash Nemotron 3.5 Lightning 30B-A3B 0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Opus 5.5 GLM-5.3-Flash

Source: Official Firecrawl Search Bench leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Tool use

All 265 in Tool use
  • AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state. Gemini 4 Neon 56.3%
  • MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks. Gemini 4 Argon 90.7%
  • BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes. Qwen3.7-Max 75%
  • τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state. Qwen 3.8 Max 55.1%
  • Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
  • DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (10 entries).
  • Pass Models scored 9 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from October 2026, for Kimi K3.1 Pro.
  • Note Frontier models Only 2 of the top 10 have a score. The strongest tested is Claude Opus 5.5 (#5).
  • Pass Headroom The best is 79%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on firecrawl.dev: 9 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 21st, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check